The genetic absence epilepsy rat from Strasbourg as a model to decipher the neuronal and network mechanisms of generalized idiopathic epilepsies.

G Model

ARTICLE IN PRESS

NSM-7241; No. of Pages 16

Journal of Neuroscience Methods xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Journal of Neuroscience Methods journal homepage: www.elsevier.com/locate/jneumeth

Invited review

The genetic absence epilepsy rat from Strasbourg as a model to decipher the neuronal and network mechanisms of generalized idiopathic epilepsies Antoine Depaulis a,b,c,∗ , Olivier David a,b , Stéphane Charpier d,e a

Inserm, U836, F-38000 Grenoble, France Univ. Grenoble Alpes, Grenoble Institut des Neurosciences, F-38000 Grenoble, France c CHU de Grenoble, Hôpital Michallon, F-38000 Grenoble, France d Brain and Spine Institute, Pitié-Salpêtrière Hospital, Paris, France e Pierre and Marie Curie University, Paris, France b

h i g h l i g h t s • • • • • •

We review several years of data on a genetic model of absence epilepsy in the rat. This model recapitulates many features of absence epilepsy and is quite predictive. It allows to record intracellular neuronal activity during spontaneous seizures. Multimodal methods showed that seizures are initiated in the somatosensory cortex. Neurons in the cortical deep layer appear to initiate spike and waves. This model allows to test new therapeutic strategies for idiopathic epilepsies.

a r t i c l e

i n f o

Article history: Received 20 March 2015 Received in revised form 28 May 2015 Accepted 28 May 2015 Available online xxx Keywords: Epilepsy Animal model Absence epilepsy Cortex Electrophysiology Magnetic resonance imaging Neural network

a b s t r a c t First characterized in 1982, the genetic absence epilepsy rat from Strasbourg (GAERS) has emerged has an animal model highly reminiscent of a specific form of idiopathic generalized epilepsy. Both its electrophysiological (spike-and-wave discharges) and behavioral (behavioral arrest) features fit well with those observed in human patients with typical absence epilepsy and required by clinicians for diagnostic purposes. In addition, its sensitivity to antiepileptic drugs closely matches what has been described in the clinic, making this model one of the most predictive. Here, we report how the GAERS, thanks to its spontaneous, highly recurrent and easily recognizable seizures on electroencephalographic recordings, allows to address several key-questions about the pathophysiology and genetics of absence epilepsy. In particular, it offers the unique possibility to explore simultaneously the neural circuits involved in the generation of seizures at different levels of integration, using multiscale methodologies, from intracellular recording to functional magnetic resonance imaging. In addition, it has recently allowed to perform proofs of concept for innovative therapeutic strategies such as responsive deep brain stimulation or synchrotron-generated irradiation based radiosurgery. © 2015 Elsevier B.V. All rights reserved.

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 What does GAERS model? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.1. The development of GAERS and its control strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.2. Why is GAERS a robust model? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00

∗ Corresponding author at: Université Joseph Fourier—Faculté de Médecine, Grenoble—Institut des Neurosciences, Centre de recherche Inserm U 836, Dynamique des Réseaux Synchrones épileptiques, Domaine de la Merci, Chemin Fortuné Ferrini, 38700 La Tronche, France. Tel.: +04 56 52 06 65; fax: +04 56 52 06 69. E-mail address: [email protected] (A. Depaulis). http://dx.doi.org/10.1016/j.jneumeth.2015.05.022 0165-0270/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: Depaulis A, et al. The genetic absence epilepsy rat from Strasbourg as a model to decipher the neuronal and network mechanisms of generalized idiopathic epilepsies. J Neurosci Methods (2015), http://dx.doi.org/10.1016/j.jneumeth.2015.05.022

G Model NSM-7241; No. of Pages 16

ARTICLE IN PRESS A. Depaulis et al. / Journal of Neuroscience Methods xxx (2015) xxx–xxx

2

3. 4.

5.

6.

The spike-and-wave discharges of GAERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.2.1. 2.2.2. Behavioral characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.2.3. Pharmacological predictivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.2.4. Ontogeny and epileptogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.3. How GAERS can help understanding mechanisms underlying epilepsy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.3.1. Genetic transmission and chromosomal mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.3.2. In vivo multi-electrodes recordings in freely movings rats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 2.3.3. An optimal in vivo preparation for real-time investigation of naturally-occurring spike-and-wave activity . . . . . . . . . . . . . . . . . . . . . . . 00 The multi-scale approach to unveil cellular basis of epileptic discharges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 How to identify the neurons generating spike-and-wave discharges? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 4.1. Functional magnetic resonance imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 4.2. Other possible uses of the GAERS model to better understand the pathophysiology of absence epilepsy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 How GAERS can help developing innovative therapeutic strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 5.1. Deep brain stimulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 5.2. Radiosurgery using synchrotron-generated microbeams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 5.3. Other possible uses of the GAERS model to develop innovative therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 General conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 00

1. Introduction Genetic animal models offer the possibility to study individuals that have a natural history close to the clinical conditions and therefore provide robust conditions to understand the pathophysiology of human diseases and their evolution throughout life. In the case of epilepsy, genetic models offer a similar ontogeny and regular occurrence of spontaneous seizures that constitute a preparation of choice and are strongly recommended by the ILAE task force on animal models of epilepsy (Simonato et al., 2014). Because most idiopathic epilepsies mainly affect children and teenagers, invasive study of their pathophysiological mechanisms cannot be conducted in the clinic for ethical reasons. Therefore, animal models are mandatory to understand these forms of epilepsy and the mechanisms underlying the generation and control of seizures. Absence epilepsy represents a prototypical form of childhood idiopathic epilepsy and different models displaying the electrical, behavioral and pharmacological characteristics of absence seizures have been developed in various species, including rodents, cats or primates by injection of pentylenetetrazol, penicillin, gamma-hydroxybutyrate or GABA agonists (see Snead, this volume). However, although these models have contributed to our understanding of absence seizure generation, the lack of recurrence and the forced induction of seizures in these preparations severely limited the study of the development of the disease, i.e., epileptogenesis. In 1982, we first reported the existence of Wistar rats with spontaneous absence seizures (Vergnes et al., 1982) and rapidly developed the Genetic Absence Epilepsy Rats from Strasbourg or GAERS, as well as a control strain. Since then, this model has been used in many studies to understand the pathophysiology of absence epilepsy and is one of the most predictive model for generalized idiopathic epilepsy. Here, we review the principal advantages of this model and present how the use of recent methodologies has allowed to better understand the genetic, cellular, network and molecular mechanisms of absence epilepsy and to develop innovative therapies. In this review a special focus is put on the methods applied to study the GAERS model. 2. What does GAERS model? Absence epilepsy is a particular epileptic syndrome where the patients show generalized non convulsive seizures characterized by a transient alteration of consciousness evidenced by a loss of responsiveness to environmental stimuli concomitant with a cessation of activity. This may be accompanied by automatisms

or moderate tonic or clonic components affecting the limbs, the eyeballs or the eyelids (Panayiotopoulos, 1999). Typical absences seizures are associated on the electroencephalogram (EEG) with bilateral, synchronous and regular 3 c/s spike-and-wave discharges (SWD) which start and end abruptly. In contrast to generalized convulsive or partial seizures, there is no postictal depression or slowing following typical absences. Absence seizures generally last 10–20 s and can occur frequently in some patients, as several hundred times per day, mainly during quiet wakefulness, inattention and at transitions between sleep and awakening. The pharmacological sensitivity of absence seizures is also specific: they are suppressed by several large-spectrum antiepileptic drugs (e.g., valproate, lamotrigine, levetiracetam) but also ethosuximide, which is ineffective in all other forms of seizures. By contrast they are aggravated by carbamazepine and phenytoine that are quite effective against generalized convulsive and partial seizures (Panayiotopoulos, 1999). Absence seizures are found in five non lesional idiopathic generalized epileptic syndromes (Porter, 1993): childhood absence epilepsy, juvenile absence epilepsy, juvenile myoclonic epilepsy, myoclonic absence epilepsy and eyelid myoclonia with absences. Beside absence seizures, these patients do not present any other neurological or neuropsychological disorders. In childhood absence epilepsy, remission is observed during adolescence in about 70% of the patients. 2.1. The development of GAERS and its control strain More than 30 years ago, we reported in Strasbourg the spontaneous occurrence of SWDs evocative of absence seizures following cortical local field potential (LFP) recordings (Vergnes et al., 1982). Breeding of selected pairs of rats displaying such pattern over 3–4 generations led us to obtain a strain with 100% of rats with SWDs. In parallel, we also bred rats from the same outbred Wistar colony in Strasbourg that were free of SWDs and over 5–6 generations we obtained the Non-Epileptic Control (NEC) strain where none of the animals display any seizures (Marescaux et al., 1992a). Both inbred strains have been maintained in Strasbourg and, since 2003, in Grenoble, as well as in Paris, Melbourne, Istanbul and Cardiff (Powell et al., 2014). The GAERS model shares a lot of similarities with another genetic model of absence epilepsy in the rat, the WAG/Rij which was inbred in the United Kingdom, then kept in Rijswijk and later at Nijmegen (The Netherlands) (Depaulis and van Luijtelaar, 2005). However, the number, cumulative total duration and mean duration of SWDs were significantly higher in GAERS compared to WAG/Rij, while the discharge frequency was higher in




3

sclerosis (Guillemain et al., 2012; Simonato et al., 2014; Depaulis and Hamelin, 2015). When asked “To your opinion, what are the most important clinical features that need to be represented in an animal model to provide the most reliable information for the clinicians on absence epilepsy (3 answers max)” the specific EEG pattern of SWDs was found as the most relevant feature to model absence epilepsy by 83% of the clinical epileptologists who answered the survey. Similarly, the pharmacological sensitivity (48%) as well as the concomitant behavior and brain structures involved (46%) also clearly emerged as critical features (Fig. 1).

Fig. 1. Key features to model absence epilepsy. Number of features selected (3 max/individual) by 82 clinical experts in epileptology to answer the question: To your opinion, what are the most important clinical features that need to be found in an animal model to be considered by clinicians as relevant to absence epilepsy? Unpublished data.

the WAG/Rij (Akman et al., 2010). Furthermore, SWDs spectra and average SWD waveforms indicated that a single cycle of the SWD contains more energy in faster components such as spike and late positive transient in the GAERS. Additionally, WAG/Rij were found to exhibit a significantly higher power between 8 and 14 Hz during the pre-SWD period. These differences in the EEGs of both animal models suggest that these variables may represent basic phenotypic features of SWDs. Nevertheless, using different approaches, from electrophysiology to molecular genetic, both models have generated an incredible amount of data during the last 30 years on the mechanisms underlying SWDs and more than 400 articles were published using either one of these two models. The present chapter focuses on the GAERS model from a methodological and translational point of view, since this strain and its control (NEC) have been explored with a large range of methods. 2.2. Why is GAERS a robust model? Although a biological model is a simplified representation of a disease, there are essential electroclinical, pharmacological and histological features that need to be represented to provide data that can be then transposed to the clinic. However, clinical questions addressed to animal models have the risk to be “lost in translation” if the model is too simplistic or represents only a limited aspect of the human disease. To limit such misunderstanding between clinicians and biologists and to ease translational research, we performed recently a survey about animal models of epilepsy by asking 6 different questions via internet (Google questionnaire) to 250 European neurologists working in university hospitals and with a recognized expertise in the treatment of epilepsies. Among them, 82 completed our questionnaire anonymously on the modeling of 4 prototypical forms of epilepsy: idiopathic epilepsy with generalized convulsive seizures, absence epilepsy, and focal epilepsies associated either with focal cortical dysplasia or hippocampal

2.2.1. The spike-and-wave discharges of GAERS SWDs is an excellent example of an EEG “signature” that can be objectively measured in several mammalian species. Indeed, SWDs have been recorded in all the genetic models that were described during the last 30 years in both mice (Noebels, 1999) and rats (Depaulis and van Luijtelaar, 2005). It is important to note that none of these models display the 3-Hz frequency of SWDs classically observed in human patients with typical absence epilepsy. In most rodent models, a frequency between 5 and 10 Hz is generally recorded (Fig. 2) and often increases with age (see below). Whether a difference in SWDs frequency compared to the human disease is a limitation of these models appears unlikely as there are many other examples where the frequency of neuronal oscillations in rodent is different than in primates (Panayiotopoulos, 1999; Depaulis and van Luijtelaar, 2005) (Porter, 1993; Bragin et al., 1999). The use of LFP recordings for models of absence epilepsy is mandatory anyway to avoid any confounding behavioral factors and to allow a precise and objective quantification of the seizures, in particular when the animal is immobilized (see below). In GAERS, SWDs start and end quite abruptly on a normal cortical desynchronized EEG background (Fig. 2). In adults, the mean frequency of spike-and-wave complexes is 8.0 ± 1.0 c/s. It is higher (9–11 Hz) during the first 1–2 s of the seizure and then slows down to 7–8 Hz for the rest of the discharge (Vergnes et al., 1982; Slaght, 2004; Marescaux et al., 1992a; Polack et al., 2007) (Figs. 2 and 6A). Their magnitude varies from 300 to 1000 ␮V depending on the brain region that is recorded (see below). When GAERS are maintained in a state of quiet wakefulness, SWDs generally last for about 25 ± 8 s, a duration that varies between individuals, since some animals can display SWDs lasting up to 60 s. They may occur about every minute, depending on the time of the day and recording conditions (Vergnes et al., 1982; Powell et al., 2014) (Marescaux et al., 1984; Depaulis and van Luijtelaar, 2005) (Marescaux et al., 1992a). Postictal depression of the LFP is never observed. Seizure characteristics have been shown recently to vary between the colonies: GAERS from Melbourne have SWDs with higher spike frequency but lower occurrence and duration than in the other colonies (Powell et al., 2014) whereas GAERS from the Grenoble colony show the greatest percentage of time spent in seizures. The identified mutation in the Cacna1h gene controlling the CaV3.2 T-type calcium channel (R1584P) was anyway present in all four GAERS colonies,

Fig. 2. The SWDs recorded in the GAERS. Color-coded time-frequency plot (bottom) of a typical SWDs spontaneously occurring in the surface EEG of the GAERS S1 in vivo (top). Unpublished data.


G Model NSM-7241; No. of Pages 16 4


but absent in all NEC rats (Powell et al., 2014) (see below). Environmental conditions may have therefore a significant impact on the expression of the SWD. As in humans, the level of vigilance plays a critical role in the occurrence of absence seizures in GAERS. When these epileptic rats are continuously recorded for 24 h, most SWDs occur during a state of passive wakefulness or drowsiness and often follow or precede slow-wave sleep (Lannes et al., 1988). On the contrary, SWDs are sporadic during active arousal, slow-wave and paradoxical sleep. In addition, a sudden noise or handling of the animals generally interrupt the on going seizure. Because of these observations, the recording conditions are critical: very few SWDs are recorded when the animals are placed in a new cage that has previously hosted a female or a mouse, or during performance of various motivated tasks (Vergnes et al., 1991). In most studies, optimal environmental conditions are usually set during the recording sessions and SWDs occurrence may be up to 1/min for about 1–2 h which is a great advantage when evaluating the effects of drugs. However, long lasting recordings revealed a lower occurrence (Saillet et al., 2013) and the existence of clusters of seizures, as in human patients. 2.2.2. Behavioral characteristics SWDs in GAERS are always concomitant with behavioral immobility and with rhythmic twitching of the vibrissae and facial muscles. Muscle tone in the neck is generally diminished and resumes at the end of the SWDs (Vergnes et al., 1982; Marescaux et al., 1984, 1992a). In some rats, light chewing with occasional tongue protrusions can be observed. Disconnection with the environment during SWDs is, however, more difficult to demonstrate in animals. When we trained GAERS to press on a lever to obtain food reward, they always interrupted their pressings during SWDs and resumed them upon cessation of the seizures (Vergnes et al., 1991). Although no behavioral impairments were initially described in GAERS (Vergnes et al., 1991), behavioral anomalies reminiscent of elevated anxiety and depression, and also psychotic-like features, were reported, at least in the Melbourne colony (Jones et al., 2008) (Bouilleret et al., 2008). It is not clear, however, if these behavioral traits are associated with the epileptic phenotype of GAERS or specific to one of the colonies of GAERS. This is to be compared with a recent clinical study showing that more than 25% of children with childhood absence epilepsy have subtle cognitive deficits, linguistic difficulties and attention deficit hyperactivity disorder and/or anxiety disorders (Caplan et al., 2008). 2.2.3. Pharmacological predictivity One of the great strength of the GAERS model is its high predictivity for antiepileptic drugs (AED). Indeed, SWDs are suppressed by all the antiepileptic drugs (AED) which are effective against human absence seizures (Panayiotopoulos, 1999), in particular ethosuximide, trimethadione, valproate, topiramate and levetiracetam. Lamotrigine, which was found less effective against absence seizures in children (Glauser et al., 2010), is also suppressive in GAERS although with a weaker effect (Grinspan, personal data; Roucard, personal data). On the contrary, AED which are either ineffective or aggravating in humans (e.g., carbamazepine, phenytoin, vigabatrin, tiagabine and gabapentine) increase SWDs in GAERS, an effect in line with what is observed in the WAG/Rij (Depaulis and van Luijtelaar, 2005). It is also interesting to note that the sensitivity of SWDs to compounds potentiating the GABAergic transmission is quite different to what is generally observed in models of other types of epilepsy. For instance, systemic injection of GABAA agonists (muscimol, THIP), GABAB agonists (R-baclofen), GABA transaminase inhibitors (gamma-vinyl GABA) or GABA reuptake inhibitors (SKF 89976 and tiagabine) induce a dose-dependent increase in the duration of SWDs (Vergnes et al., 1984; Marescaux et al., 1992b), in line with the aggravating effects of vigabatrin

reported in clinical studies (Parker et al., 1998) and recent data on the tonic role of GABA in this model (Cope et al., 2009). 2.2.4. Ontogeny and epileptogenesis Childhood absence epilepsy is known to be associated with brain development and maturation (Crunelli and Leresche, 2002). In humans, typical childhood absence epilepsy occurs between the age of 2 and 8 years and remission takes place around puberty in most patients (Loiseau et al., 1995). In GAERS, SWDs were initially first detected around 30–40 days of age (Vergnes et al., 1986). More recently, LFP recordings performed in the site of seizure initiation, the somatosensory cortex (see below), revealed that abnormal oscillations first occur as soon as 14 days post natal (P14) and progressively evolve as SWDs with spike-and-wave patterns emerging at P25 and then increasing in number with age during a discharge (Girod et al., 2012) (Jarre et al., in prep). The duration, occurrence and frequency of the “immature” and then mature SWDs increase with age. Their number reaches a plateau around 4 months and SWDs can be recorded until the death of the animals (Vergnes et al., 1986). This persistence of SWDs in GAERS, as well as in the other rodent models of absence, is in contrast with the evolution of childhood absence epilepsy in humans. Whether this difference indicates that GAERS models a form of absence epilepsy that persists in adulthood or that mechanisms underlying their remission is lacking in rodents, remains to be examined. 2.3. How GAERS can help understanding mechanisms underlying epilepsy Mainly because of its genetic origin but also due to the high recurrence of SWDs and the ease to identify them, the GAERS model has become one of the most predictive model to study the pathophysiology of a neurological disease. Several translational studies have therefore been possible thanks to this model in particular for our understanding of the etiology of SWDs and the neural circuits that generate them. Here, we illustrate this point with few recent studies from our groups and others which have used the GAERS model. 2.3.1. Genetic transmission and chromosomal mapping The GAERS, as well as the NEC, are fully inbred rats, i.e. they are homozygous for all autosomal genes (Marescaux et al., 1992a). Breeding of selected parents from the initial colony in Strasbourg over 3–4 generations produced a strain in which 100% of the rats were affected. In epileptic × non epileptic F1 generation, more than 95% of the animals showed SWDs after 6 months, suggesting a dominant transmission. Similar SWDs were recorded in males and females, indicating that the transmission is autosomal. Interindividual variability suggests that the inheritance of SWDs is not due to a single gene locus and/or that environmental effects might play a role. This mode of inheritance was confirmed in F2 (F1 × F1) and backcross (F1 × control) generations (Marescaux et al., 1992a). When F2 population was generated by breeding GAERS with Brown Norway rats, a polygenic inheritance of SWD-related phenotypes was shown and three quantitative trait loci were identified that could control different variables of SWDs (e.g., frequency, amplitude, duration). In this study, the age of the animals was found to be a major factor influencing the detection of genetic linkage to the various components of the SWDs (Rudolf et al., 2004). The development of two inbred strains from the same initial colony has appeared as a very powerful tool to study the possible mutations involved in a genetically complex idiopathic epilepsy (Frankel, 2009). Indeed, a functional mutation in the Cacna1h gene encoding the Cav3.2 low-voltage activated Ca2+ channel was found using the two strains (Powell et al., 2009). In addition, this study showed that the effect was due to a gain-of-function splice




5

Fig. 3. Primary somatosensory cortex as the trigger zone for SWDs in the GAERS. (A) In order to elucidate the initiation site of SWDs in the GAERS, simultaneous EEG recordings were made in freely-moving animals at the indicated sites (mm from the bregma) in the somatosensory cortex (red points), motor cortex (blue points) and ventrobasal (VB) complex of the thalamus (green points). (B) The corresponding records (as indicated by the numbers) clearly show that SWDs begin in the primary somatosensory cortex before propagating to motor cortex and thalamus. During the same EEG recording sessions, short spike-and-wave activity occurred in the primary somatosensory cortex without concomitant paroxysm in motor cortex and thalamus (right). (A) and (B) modified from (Polack et al., 2007). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

variant mutation, and was semi-dominant, explaining about 30% of the phenotypic variance in the cross. In heterologous expression studies, it was shown that the GAERS splice variant allele on Cav3.2 conferred faster recovery from channel inactivation and greater charge transference during high-frequency bursts. This is in agreement with a previous study that showed a selective increase in the T-type conductance in GAERS nRT neurons (Tsakiridou et al., 1995). It is also in line with the role of the low voltage activated Ca2+ channel in thalamic burst firing (Crunelli and Leresche, 2002) and genetic data in human patients (Hughes, 2008). Whether this model will help for the discovery of new mutations involved in absence epilepsy remains to be confirmed since it is very likely that different combinations of gene mutations may lead to SWD. 2.3.2. In vivo multi-electrodes recordings in freely movings rats The possibility to record seizures regularly in either freely moving or immobilized GAERS (see below) has allowed to confirm and further describe the region of the brain where absence seizures are initiated. The initial work on the WAG/Rij model first suggested that SWDs were initiated in the somatosensory cortex (Meeren et al., 2002, 2005). Since then, we explored this region of the cortex which was poorly investigated in our initial mapping study (Vergnes et al., 1990). Using multielectrode LFP recordings in freely moving adult GAERS, we recorded simultaneously sites located in the motor and somatosensory cortices, as well as in the ventrobasal thalamus (Fig. 3A) and observed that 91.9% of SWDs first occurred at sites located in the primary somatosensory cortex (S1) (Fig. 3B). SWDs could begin under any electrodes located in this cortical region but preferentially from more rostral sites. Only 8.1% of the discharges started simultaneously at all cortical sites but no SWDs ever began in the motor cortex or in the thalamus in this study (Polack et al., 2007). A consistent delay of about 1 s was observed between the beginning of the SWDs in the somatosensory cortex and their beginning in the motor cortex and of about 1.3 s between the somatosensory cortex and the thalamus. Measurements of

association strength between field potentials recorded at different cortical and thalamic sites in freely moving GAERS also suggested that SWDs originate from the facial somatosensory cortex. In addition, we found that short (1–2 s) SWDs could be recorded in this region that did not spread to other regions whereas the reverse was never observed (Polack et al., 2007) (Fig. 3B, right pannel). These short and localized SWDs were not associated with chewing or twitching of the vibrissae. We later confirmed the specific role of the S1 cortex in the initiation of SWDs by multielectrode array recordings (64-electrode FlexMEA) and quantification of the nonlinear correlation coefficient h2 as a function of a time shift between two records (Lopes da Silva et al., 1989) in immobilized GAERS (Pouyatos, unpublished data). Altogether, these findings demonstrated an initiation of SWDs in the facial somatosensory cortex also called the barrel field cortex and an intracortical propagation of ictal activity as the mechanism of primary generalization of spontaneous SWD, in agreement with the initial study in the WAG/Rij rat (Meeren et al., 2002, 2005). The concept of a specific cortical region initiating SWDs is also in line with recent studies in human patients that revealed the critical role of the frontal cortex (Holmes et al., 2004; Sadleir et al., 2006). Whether similar mechanisms are involved in the generation of SWDs in human patients remain to be established. 2.3.3. An optimal in vivo preparation for real-time investigation of naturally-occurring spike-and-wave activity Many experimental approaches can be used to investigate the multiple facets of epileptic processes. In vitro or ex-vivo experiments (Avanzini et al., 2007) are necessary for the elucidation of molecular alterations of ion channels or neurotransmitter signaling but preclude the “spontaneous” occurrence of seizures. By contrast, the local and long-distance synchronizing mechanisms require in vivo preparations preserving the whole brain connectivity (Timofeev and Steriade, 2004) (Polack et al., 2009). All these approaches are complementary and provide crucial information.




Fig. 4. Experimental design used for a multi-scale analysis of spike-and-wave activity in the GAERS in vivo. (A) To elucidate the neuronal basis, the initiation processes and the mechanisms of large-scale synchronization of SWD, multi-site and multi-scale electrophysiological investigations were made in the GAERS in vivo. The red dots on the schematized representation of the GAERS head indicate epidural EEG recording sites from sedated-immobilized GAERS, including primary somatosensory cortices (S1) and the facial primary motor cortex (M1). Methodical exploration of the intracellular (indicated by the microelectrodes) activities of the different neuronal elements of the S1 region (P, pyramidal neurons; I, interneurons; S, stellate neurons localized in the indicated cortical layers) and related thalamo-cortical (TC) nuclei (VPm, ventro-postero medial; POm, posterior medial) (enlarged networks at right) was made to identify the neurons underlying the initiation of SWD. In all experiments, intracellular recordings were performed concomitantly with local and/or distant EEG or local field potential (LFP) records. The participation of the GABAergic neurons of the nucleus reticularis thalami (NRT), which inhibit TC neurons, was also investigated by the means of intra- and extracellular recordings. The glutamatergic (Glu) and GABAergic (GABA) synapses are indicated in grey and black, respectively. The involvement of cortical and thalamic neurons in the initiation of seizures was notably assessed by the impact of their specific inactivation via local microinfusion of tetrodotoxin (TTX) (see Fig. 6D). This in vivo preparation also allows activating the sensory pathway, from whiskers to S1 neurons by natural stimulations (whiskers deflection induced by air-puff). The reciprocal connections depicted between POm and nRT neurons also exist for VPm neurons. (B) These in vivo experiments, associating EEG, intra- and extracellular recordings together with pharmacological manipulations were made in GAERS sedated with fentanyl, a potent narcotic analgesic, allowing the investigation of spontaneously occurring SWDs and of their cellular correlates.

However, the multi-dimensional defects supporting seizure activity are causally intricate and cooperate to initiate, and eventually generalize, epileptic paroxysms. Therefore, the most promising strategy to unveil the brain events operating during seizures is to monitor simultaneously the neural network involved in the seizures along with the synaptic and membrane properties of individual neurons (Fig. 4A). This approach, which represents a technical challenge in living preparations, has been successfully applied to the GAERS to elucidate the processes of initiation and generalization of genetically-determined SWDs. The crucial point was to develop an acute in vivo preparation allowing the spontaneous occurrence of SWDs together with the required conditions to perform concomitantly EEG and stable intracellular recordings. The neuroleptanalgesia was first used for this purpose (Pinault et al., 1998; Charpier et al., 1999; Slaght et al., 2002). This procedure involves combined systemic injections of a major tranquilizer, usually a neuroleptic, such as droperidol or haloperidol, with a potent synthetic opioid, the fentanyl (Fig. 4B), to produce a detached, pain-free state of immobilization (Flacknell, 1996). Erratic spontaneous muscular contractions can however occur in the fentanyl-sedated rodents (Simons and Carvell, 1989)

making intracellular recordings unsecure, occasionally causing serious damages of impaled neurons. Thus, neuroleptanalgesia was completed by injection of neuromuscular blockers to prevent any active movements and GAERS were artificially ventilated (Vergnes et al., 1990) (Charpier et al., 1999) (Pinault et al., 1998; Slaght et al., 2002). Because dopaminergic system can alter spike-andwave activity in the GAERS, via modulation of GABAergic synaptic inhibition of thalamo-cortical neurons (Yagüe et al., 2013) or by controlling the activity of basal ganglia circuits ((Deransart et al., 1998; Deransart and Depaulis, 2002), the neuroleptic agent is no more required for the in vivo preparation of GAERS and the fentanyl can be used alone allowing a robust analgesied, narcotized and sedated state in curarized animals (Polack et al., 2007, 2009; Chipaux et al., 2011). The fentanyl-curare preparation, now routinely used to record neuronal activity in the GAERS (see below), offers many advantages to investigate SWDs in a condition as close as possible to their “natural” occurrence. First, this preparation does not alter the temporal and spatial properties of cortical SWDs compared to those recorded in freely moving GAERS, including their internal frequency, duration, interval of recurrence, voltage waveform and




7

Fig. 5. Cellular basis of the cortical spike-and-wave complex. (A) In vivo experimental arrangement. The neuronal activity underlying cortical SWDs in the GAERS is investigated by concomitant recordings of surface EEG and intracellular (Intra) activities of pyramidal neurons and interneurons. (B) Intracellular activity of a layer 5 pyramidal neuron (bottom record) simultaneously recorded with the surface EEG (top record). The occurrence of an SWDs in the EEG is accompanied in the neuron by rhythmic depolarizations, which are superimposed on a tonic membrane hyperpolarization (dashed line). (C) Segment (0.5 s duration) of the paired recording shown in (A) during the ictal period. The spike (S) and wave (W) components of the SWDs and their intracellular correlates are delimitated by the grey and red box, respectively. The dashed line indicates the mean value of interictal membrane potential. The brisk firing of the neuron associated with the S-component is interrupted by hyperpolarizing synaptic events (oblique arrows). In this cell, the depolarizing phase of the oscillatory cycle displays a ramp-like profile (crosses arrow). (D) Alignment of neuronal activities associated with the S-component. The depolarization (DC injection of +1 nA) of a pyramidal neuron (recorded with a KAc electrode, top record) above equilibrium potential of Cl− shows an early synaptic depolarization (arrow) rapidly interrupted by a brisk hyperpolarization. When the neuron is loaded with Cl− (recording with a KCl electrode, middle record), the synaptic hyperpolarization is converted into a large depolarizing potential, suggesting the presence of a GABAA receptor-mediated synaptic conductance. The synaptic inhibition of pyramidal neurons during the S-component is temporally correlated with bursting activity in cortical interneurons (Intern, lower record). (E) Continuous intracellular record from a cortical neuron (Intra) showing the voltage responses to injection of hyperpolarizing current pulses of constant amplitude (−0.8 nA), before and during a SWDs (EEG). The averaging (8 or more successive trials) of the cell responses (bottom) clearly shows an increase in membrane resistance during SWDs. (B) and (C) modified from (Polack et al., 2007). (D) and (E) modified from Chipaux et al. (2011) and Charpier et al. (1999), respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

site of initiation. This was evidenced in a study combining multisite EEG records from freely moving and sedated-immobilized GAERS (Polack et al., 2007). Second, the rodent mean arterial blood pressure ranges during fentanyl sedation is similar to healthy awake rats (Bruno and Sakmann, 2006) and the related baseline cortical EEG patterns are characterized by fast (up to 50 Hz) waves of small amplitude comparable to the “activated” waking pattern (Constantinople and Bruno, 2011) (Altwegg-Boussac et al., 2014). Third, neocortical neurons under fentanyl displayed low values of membrane input resistance (Altwegg-Boussac et al., 2014), consistent with the “high-conductance state” characteristic of these cells during wakefulness (Destexhe et al., 2003), without altering the other membrane properties (Mahon et al., 2001). Finally, it is important to mention that fentanyl has minor impact on the functional and electrophysiological properties of the primary somatosensory cortical pyramidal neurons. Indeed, fentanyl acts on the opioid receptors mu (Inoue et al., 1994), which are particularly sparse in the S1 cortex, including the barrel cortex, being most densely concentrated in the limbic and frontal areas (Sahin et al., 1992). Moreover, sensory-evoked responses of barrel cortex neurons and related thalamo-cortical circuits, recorded from normal rats and GAERS, are not different between fentanyl sedation and wakefulness (Simons et al., 2007; Chipaux et al., 2013). GAERS under fentanyl sedation, combined with paralyzing agents, therefore appear as an optimal preparation to investigate in vivo and in real-time the cellular basis of SWD. Following, we

described recent advances, part of them unexpected, that were obtained in the GAERS model by the means of this preparation. 3. The multi-scale approach to unveil cellular basis of epileptic discharges The use of in vivo preparation of GAERS has allowed for important insights in the understanding of neuronal and network mechanisms underlying genetically-determined SWDs. In particular, the simultaneous recordings of EEG and intracellular activities (Fig. 5A) provided a powerful tool to unveil the cellular basis of spontaneously occurring epileptic discharges. Indeed, although paroxysmal EEG waveforms, as recorded from the surface of the brain, describe the electrical features of epileptic seizures (the so-called “spike-and-wave” pattern in the case of absence epilepsy), they do not provide any information on the underlying neuronal activities, the neuronal types involved, their firing patterns or the synaptic and membrane properties cooperating to produce paroxysmal discharges and their synchronization. To the best of our knowledge, paired recordings in the GAERS of surface EEG and corresponding intracellular activities (Fig. 5A) allowed for the first time the identification of the neuronal correlates of idiopathic generalized seizures. The occurrence of spike-andwave activity on the surface EEG of S1 cortex is correlated in the corresponding deep-layer pyramidal neurons with suprathreshold rhythmic depolarizations, superimposed on a tonic membrane




hyperpolarization (Fig. 5B) that lasts throughout the seizure (Polack et al., 2007, 2009; Polack and Charpier, 2009; Chipaux et al., 2011, 2013). This observation was rather unexpected as it indicates that genetically-determined SWDs are associated with an increase in cell polarization, a steady change in neuronal state that drastically differs from the classical sustained depolarizing shift found in other forms of seizures, using the same fentanyl-curare preparation (Langlois et al., 2010). Thus, in contrast with a general assumption (Gorji and Speckmann, 2009), paroxysmal depolarization shift is not an ubiquitous characteristic of epileptiform activity of individual neurons. Moreover, the mean firing rate of GAERS cortical neurons during SWDs is not increased compared to the preceding quiescent periods, even in the ictogenic neurons (Polack et al., 2007, 2009). This other unforeseen finding indicates that cortical neurons during seizures do not systematically generate “paroxysmal” discharges and that seizures, at least those associated with absences, are rather associated with a change in the temporal pattern of neuronal firing. A detailed analysis of the intracellular correlates of SWDs revealed the origin of the two waveforms that compose SWD: the “spike” (S) and the “wave” (W) components (Fig. 5C, top record). Consistent with the dipole theory for electrogenesis of spontaneous EEG signals (de Munck et al., 1992), the S-component is systematically associated with a large, sinusoid-like, membrane depolarization in cortical neurons, which can generate a brisk firing (Polack et al., 2007, 2009) (Fig. 2C). These rhythmic depolarizations are sculpted by the summation of sharp depolarizing events, their amplitude increases as a function of the membrane polarization while their frequency of recurrence remains unaffected (Slaght et al., 2002; Polack et al., 2007; Chipaux et al., 2011). Altogether, these properties indicate that the neuronal activity underlying the S-component results, at least in part, from synaptic potentials even though voltage- and time-dependent membrane processes cannot be excluded. How to dissect the synaptic components composing the rhythmic depolarizations underlying the S-component? Although elucidating such complex processes is highly challenging from living animals, we used different strategies in our in vivo preparation of GAERS. In addition to the properties of cell oscillations described above, which strongly suggest the participation of a powerful excitatory synaptic drive, likely glutamatergic, the firing of cortical neurons correlated with the S-component is interrupted by upward and downward membrane potential fluctuations (Fig. 5C, arrows) resembling a mixture of excitatory and inhibitory synaptic events (Polack et al., 2007; Chipaux et al., 2011). How to make sure that an inhibitory component also participates to the paroxysmal depolarization and how to determine its origin? In vivo intracellular recordings offer many possibilities to answer these questions. First, a Cl− -dependent, presumably GABAergic, synaptic potential can be identified on the basis of its voltage-dependency, being depolarizing or hyperpolarizing when the membrane potential of neuron is displaced below and above the reversal potential of Cl− , respectively (Eccles, 1964). This procedure was successfully applied to the cortical neurons of GAERS during seizures, revealing during the S-component a succession of synaptic events composed by an initial short depolarization (likely glutamatergic) abruptly followed by a large potential that reverses in polarity as expected from a Cl− -dependent synaptic event mediated by GABAA receptors (Chipaux et al., 2011) (Contreras et al., 1997) (Fig. 5D, top record). Second, the presence of Cl− -dependent membrane conductance can be confirmed by the induction of a large outward flow of negative ions–thus depolarizing–when the neuron is recorded with a KCl electrode, a procedure that overloads the cell with the anion and thus reverses its driving force. Consistently, Cl− -loaded cortical neurons exhibit a large additional membrane depolarization correlated with the S-component (Fig. 5D, middle record), demonstrating the presence of a Cl− -dependent synaptic potential

(Chipaux et al., 2011). Finally, because GABAergic inputs to cortical pyramidal neurons quasi exclusively originate from local inhibitory interneurons (Fig. 4A), the presumed GABAA synaptic events should be correlated with the activation of GABAergic interneurons connecting pyramidal cells. In vivo intracellular recordings of GAERS cortical interneurons, identified on the basis of their morphology and distinctive electrophysiological features, confirm this hypothesis. Inhibitory cells display bursting activity that is time-locked with the S-component in the EEG and temporally correlated with the Cl− -mediated synaptic inhibition in the neighboring pyramidal neurons (Chipaux et al., 2011) (Fig. 5D, lower record). In contrast, the W-element is systematically associated with a membrane hyperpolarization and a neuronal electrical silence (Fig. 5C, red box), which delineate the sustained polarization of neurons during the seizure (Fig. 5B). This could reflect a synaptic disfacilitation, i.e. a transient interruption of the tonic excitatory synaptic drive (as seen between seizures), leading the cell toward its resting potential and causing an increase in membrane input resistance due to the break of synaptic conductance (Charpier et al., 1999) (Fig. 5E). During this quiescent period, some pyramidal cortical neurons exhibit a “ramp-like” depolarization in the early phase of the oscillatory cycle (Fig. 5C, crossed arrow), which may result from a depolarizing mixed cationic conductance activated by the hyperpolarization (Ih ). This voltage-gated current, possibly activated during the W-component, could act as a pacemaker-like process boosting the rhythmicity of pyramidal neurons during seizures (He et al., 2014). The experimental strategy described above, combining in vivo EEG recording with intracellular recordings of pyramidal neurons and interneurons, together with manipulation of membrane potential and internal ion concentration of recorded cells, help us to understand the successive neuronal events underlying the spikeand-wave complex recorded at the surface of the brain. Indeed, the S-component of the cortical paroxysm derives from the synchronization of large neuronal depolarizations composed by an initial excitatory synaptic potential, likely due to glutamatergic interactions between pyramidal neurons, possibly amplified by persistent voltage-gated Na+ current (Klein, 2004; Chipaux et al., 2011). This excitation, generating action potentials and driving the membrane potential of pyramidal cells above the reversal potential of Cl− , is rapidly followed by a burst of hyperpolarizing synaptic events, due to the recurrent activation of interneurons, limiting the firing of pyramidal cells (Chipaux et al., 2011). The consecutive interruption of ongoing synaptic activity in the cortical network, transiently keeping the neurons in a state of electrical silence, explains the W-component on the EEG. 4. How to identify the neurons generating spike-and-wave discharges? One of the most delicate issues of the research in epilepsy, especially for generalized seizures, is to determine the neural network and, ideally, the neurons (“ictogenic neurons”), that initiate the epileptic discharge. In the case of absence seizures, the challenge is to identify neurons that: (i) are the first to be activated during each SWD, (ii) drive directly or indirectly other neurons, (iii) are specifically required for the occurrence of generalized SWDs and, (iv) display pro-epileptogenic properties. Moreover, it is also expected that such seizure-initiating neurons present some morphological defects and provide a potent target for anti-absence drugs. All these criteria have been met in pyramidal neurons of the deep-layers of the primary somatosensory cortex (S1) of GAERS. Following the work performed in freely moving GAERS (see above), a successful identification of ictogenic neurons was obtained by the combination of complementary experimental approaches in the sedated-immobilized GAERS. As shown by the




9

Fig. 6. Multiple approaches are required to identify ictogenic neurons. (A) Examples of preictal and ictal EEG activities simultaneously recorded from S1 (top record) and the primary motor cortex (M1, bottom record) of freely moving GAERS and corresponding averaged time–frequency maps (4 GAERS). Note that paroxysmal oscillations first occur in S1, with the specific presence of 10–11 Hz oscillations (arrows) in the ictogenic cortex just before the generalized SWDs (vertical line). (B) Superimposition of the intracellular activities (bottom traces) recorded in a layer 2/3 (green), layer 4 (blue), and layer 5/6 neuron (red) and the corresponding spike-wave complexes (top, black records). Intracellular and EEG records are aligned using the peak negativity of the EEG S-component as the zero-time reference. (C) Temporal evolution of the relationships between S1 EEG and intracellular activity of a neuron of the ventro-postero medial (VPm) thalamic nucleus (top records). The nonlinear correlation indices h2 indicate that the strength of association between the cortex and thalamus gradually increases throughout the seizure. The corresponding time delays shows unidirectional coupling, from the cortex to thalamus, during the first few seconds of the seizure, then temporal relations fluctuate. (D) S1 is required for cortical SWD. (D1) Application of TTX on S1 (ipsi), which interrupts firing activity in the corresponding deep-layer pyramidal neurons (lowest record), suppresses SWDs in the ispilateral S1, ipsilateral M1 and contralateral S1 EEGs (top records), this later still displaying some residual oscillations (arrow). (D2) When TTX is applied on M1, despite the silencing of M1 pyramidal neurons (lower record) and the disappearance of SWDs in the corresponding EEG (M1 ipsi), the ipsilateral S1 EEG still produces SWDs. The insets in (D1) and (D2) show active periods of the neurons recorded before application of TTX. Their action potentials are truncated. (E) Ethosuximide (ETX) normalizes the activity of S1 pyramidal neurons of the GAERS. Each paired records shows the EEG and the corresponding intracellular activities recorded in a layer 5 neuron from GAERS S1, before (GAERS) and 30 min after ETX injection (GAERS + ETX), and from a layer 5 cortical neuron located in the S1 of a non-epileptic rat (Non-epi). (A) and (B) modified from Polack et al. (2007). (C) and (D) modified from Polack et al. (2009) and (E) from Polack and Charpier (2009). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

averaged time–frequency power analysis of SWDs recorded in the S1 cortex in freely moving GAERS (Fig. 6A, top), the transition from interictal to ictal activities is accompanied by a transient increase of power in the 9–11 Hz frequency band, which is specific to this cortical area (Fig. 6A, bottom). To further elucidate the process of seizure initiation within the S1 cortex, it was essential to analyze the temporal sequence of firing in neurons located in the different cortical layers (Figs. 4A and 6B). For this purpose, the timing of individual action potentials was measured in the different layers relative to the peak negativity of the corresponding local EEG. The S-component was chosen as the zero-time reference (Fig. 6B, top record) because it provides a reliable and reproducible marker of synaptic synchronization in the local cortical network (see above). The firing of layer 5–6 neurons systematically precede, throughout SWD, the discharge of neurons in layers 2, 3 and 4, suggesting a leading role of deep-layer neurons in the generation of SWDs within the S1 cortex (Fig. 3B). Because S1 cortex receives excitatory synaptic inputs from specific thalamo-cortical (TC) nuclei, chiefly the ventro-postero medial (VPm) and posterior medial (POm) nuclei (Fig. 4A) (Feldmeyer et al., 2013), it was essential to assess the directionality of TCS1 cortex interactions during seizures. We obtained consistent

findings, demonstrating the leading role of S1 on the related TC neurons. First, using the same temporal reference as for cortical interactions (Fig. 6B), action potentials generated in both groups of thalamic cells during the seizures followed the firing of the related pyramidal neurons in S1 cortex (Polack et al., 2009). Second, the directionality of information flow between thalamus and S1 was further examined by measuring concomitantly the degree of statistical association and the time delay between S1 EEG and thalamic intracellular activities, from VPM and POm neurons (Fig. 6C, top records). This was done by quantification of the nonlinear correlation coefficient h2 as a function of a time shift between the two records (Lopes da Silva et al., 1989). Although classical methods such as cross-correlation or coherence yield reliable results only for linearly related signals, this nonlinear index can reveal signal interdependencies under more general conditions, including seizures activities (Meeren et al., 2002). h2 Values between thalamic cells and cortical EEG gradually increase just prior to SWDs and are maintained at an elevated level during the seizure (Fig. 6C, middle panel). Oscillatory activity in VPM and POm thalamic cells lagged spike-and-wave activity in S1 cortex during the first second of the discharge, and then no preferential directional coupling was evident during the later parts of the SWD. These




findings indicate that cortex leads thalamus during the onset of the seizure activity and suggest a sustained interplay between the two structures to maintain coherent cortico–thalamo-cortical oscillations during SWDs (Meeren et al., 2002; Polack et al., 2009). However, it seems that the different thalamocortical systems do no play an equivalent role. Indeed, although the “specific” thalamic relay neurons display a relatively low firing rate during seizures (Pinault, 1996; Paz et al., 2007; Polack et al., 2009), neurons of the “non-specific” thalamic nuclei, including midline (Paz et al., 2007) and intralaminar (Seidenbecher and Pape, 2001) thalamocortical neurons, exhibit synchronized high-frequency burst-like activity in conjunction with the S-component of the cortical SWD, suggesting a major role of the “non-specific” thalamic system in the synchronized oscillations within the cortico-thalamo-cortical loop. Among the criteria required to demonstrate a causal role of S1 neurons in the initiation of absence seizure is their specific requirement for the appearance of SWD. This was evidenced in vivo by pharmacological inactivation of various cortical and thalamic regions. Indeed, if S1 neurons are causally and specifically implicated in the generation of SWDs, their inactivation must preclude the appearance of seizures in S1 as well as in other parts of the brain. Conversely, inactivation of distant brain areas – even those connected with S1 – will not prevent the occurrence of paroxysmal oscillations in this part of the cortex. Consistent with these predictions, local infusion of tetrodotoxin (TTX) in S1 cortex, a blocker of voltage-gated Na+ channels that prevents action potentials discharge, interrupt the firing and synaptic activities in S1 neurons (Fig. 6D1, lower records) together with the disappearance of SWDs in all EEG records (Fig. 6D1, upper records), only residual oscillations being still present in the contralateral S1 EEG (Fig. 6D1, arrow) (Polack et al., 2009). On the opposite, the same pharmacological blockade of other cortical (Fig. 6D2, lower records) and thalamocortical neurons suppresses local paroxysmal oscillation but has no effect on the ability of S1 to produce SWDs (Fig. 6D2) (Polack et al., 2009). The demonstration of the causal action of S1 pyramidal neurons in the initiation and the generalization of SWDs makes these neurons a potent target for anti-absence medicine. Many in vivo pharmacological investigations in the GAERS support this hypothesis. First, local application of ethosuximide (ETX) in the S1 cortex of the GAERS specifically suppresses generalized SWDs (Manning et al., 2004), suggesting that the anti-absence action of ETX is due to a direct and preferential effect on S1 neurons. Second, a systemic injection of ETX in the GAERS, at therapeutic doses, converts the hyperactive and hyperexcitable S1 pyramidal neurons (Polack et al., 2007, 2009) (Chipaux et al., 2011, 2013) into “normal” neurons, including physiological firing rate, membrane potential values and current-evoked responses, similar to those recorded in the homologous pyramidal neurons in the S1 of non-epileptic rats (Fig. 6E). The findings briefly described above strongly support the hypothesis that SWDs in the GAERS primarily originate from aberrant activity of pyramidal neurons located in the deep layers of S1. This activity associates an excessively depolarized membrane potential together with an elevated and highly regular spontaneous firing during interictal periods, with rhythmic brisk discharges during SWDs that exceed and precede the firing of distant cortical and thalamic neurons. However, it is also plausible that specific molecular alterations in GAERS thalamocortical neurons, affecting their excitability and their oscillatory patterns, may have a dual role in seizure activity. For instance, in vitro experiments suggest that abnormal functions of pacemaker currents (Ih ), carried by hyperpolarization-activated cation non-selective (HCN) channels, could promote epileptogenesis in GAERS and, by compensatory mechanisms stabilizing Ih function, may also contribute to the termination of thalamic oscillations (Kuisle et al., 2006).

4.1. Functional magnetic resonance imaging Distinguishing neural drivers from other brain regions is also an essential question to understand the mechanisms involved in SWDs and more generally in pathological activities of CNS disease. To this aim, sophisticated signal analysis techniques have been developed to estimate the connectivity between distant regions based on electrical brain signals (David et al., 2004; Quian Quiroga et al., 2002). However, although the time resolution of these techniques is quite satisfactory, the spatial resolution is unsufficient, even with the use of multielectrode matrix, because of the impossibility to have access to all parts of the brain simultaneouly. The hypothesis that SWDs originate from a specific brain region thus needed to be further examined using an approach with both excellent spatial coverage and anatomical resolution, such as functional magnetic resonance imaging (fMRI). During neuronal activation, fMRI is sensitive mainly to changes of local perfusion and oxygen uptake by neurones (Rasch et al., 2009) and therefore provides an indirect measure of neuronal activity. Nowadays, concurrent EEG and fMRI recordings have become the method of choice for SWDs in patients (Hamandi et al., 2008; Carney et al., 2012; Moeller et al., 2010). The possibility to use an in vivo preparation that allows the perfect immobilization of the animals (see above) was a great advantage to perform such an approach in GAERS. The main added-value of studying SWDs in GAERS with fMRI, as compared to clinical studies, is the possibility to use a contrast agent in animals that dramatically increase the signal-to-noise ratio. In addition, the experimental session in animals can last much longer than in patients, and thus it is easier to look at the reproducibility of fMRI activations and to obtain significant activations. To perform fMRI during seizures in GAERS (David et al., 2008; David, 2011), we equipped the animals under ketamine with carbon electrodes located on the skull near the midline, several hours prior to the experiments. Two additional carbon electrodes were used to monitor cardiac activity. The rats were first anaesthetized with isoflurane, their femoral artery was catheterized to allow administration of an iron-based superparamagnetic contrast agent as well as infusion of curare and analgesics. Neuroleptanalgesia was induced as described above, a tracheotomy was performed and animals were ventilated during the rest of the experiment. This allowed to secure the animals in a magnetic resonance-compatible, customized, stereotaxic headset with ear and tooth bars, for up to 4 h in the magnet. Magnetic resonance imaging was performed in a horizontal bore 2.35 T magnet (Bruker Spectrospin, Wissembourg, France), equipped with actively shielded magnetic field gradient coils (Magnex Scientific Ltd., Abdington, UK) and interfaced to a SMIS console (SMIS Ltd, Guildford, UK). A linear volume coil was used for excitation and a surface coil was used for detection (Rapid Biomedical GmbH, Rimpar, Germany). EEG and cardiac signals were sampled simultaneously with fMRI and EEG and fMRI temporal coregistration was ensured by the EEG acquisition software recording a TTL signal from the magnetic resonance system at each volume acquisition (David et al., 2008). The analysis of our fMRI data involved three distinct components (David et al., 2008). First, we characterized the hemodynamic response to seizure activity using conventional statistical parametric mapping to identify regionally specific responses. We then characterized the regional variations in the hemodynamic responses by optimizing the parameters of a hemodynamic model for different regions of interest (ROI) separately. The second component of our analyses comprised a comparative evaluation of vector regression models (Granger causality) (Goebel et al., 2003) and Dynamic Causal Modelling (DCM) (Friston et al., 2003). Granger causality tests for statistical dependencies among observed (time-lagged) physiological responses, irrespective of how they are caused whereas DCM represents hidden states that




11

Fig. 7. Functional MRI of SWD. (A) fMRI group analysis (n = 8) showing regions with significant changes of cerebral blood flow (CBV) during SWDs (red: increase of CBV; blue: decrease of CBV). 3D views of a canonical rat brain “cut” at different positions in the anterior (A)–posterior (P) axis. Local field potential (LFP) recordings of a typical SWDs are shown for four regions of this network, with a clear start of SWDs in the first somatosensory cortex (S1). (B) Comparison of CBV dynamics in S1 and thalamus. Though LFP dynamical analysis clearly indicated the precedence of S1 over the thalamus (“LFP” arrow), CBV changes in S1 are so abnormally slow that there is an inversion of the order of the observed hemodynamics between the two regions (the figure shows a SWD selected for its very long duration, i.e. 280 s, in order to clearly show the transition of hemodynamics between baseline and ictal state). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cause the observed data and are therefore causal models in a true sense. Finally, we used two complementary approaches for cross-validation based in depth electrode recordings: (i) a simple characterization of propagation delays of SWDs to establish the direction of connections through temporal precedence and (ii) asymmetries in directed generalized synchrony of SWDs time series. Using these 3 complementary approaches, we obtained the following data (Fig. 7): 1. An increase of cerebral blood volume (CBV) signal during SWDs was found in the barrel field of the primary somatosensory cortex (S1BF), the centromedial, mediodorsal and ventrolateral parts of the thalamus, the retrosplenial cortex, and the reticular part of the substantia nigra. All these structures have been shown to be involved in the generation, spreading or control of SWDs in GAERS (Depaulis and van Luijtelaar, 2005). The cerebellum and nuclei of the pons and of the medulla oblongata were also found activated. In addition, several areas showed a decrease of CBV signal, such as the striatum, the limb representation of the primary somatosensory cortex, the visual cortex and the secondary motor cortex. 2. The hemodynamic response functions were found to last significantly longer in S1BF than in the other regions. Using biophysical modeling, we explained this slow dynamics by the near suppression of the autoregulation mechanisms of cerebral blood flow (CBF) on vasodilatation. Functional hyperemia, which matches the delivery of blood flow to the activity level of each brain region, requires coordinated cellular events that involve neurons, astrocytes, and vascular cells (Iadecola and Nedergaard,

2007). Deregulation of the function of any of these cell types in S1BF thus appears as a plausible physiological mechanism to explain the abnormally long time constant of CBF feedback that we found. 3. Granger causality at the group level disclosed the predicted architecture, in which S1BF drove the other regions, only when applied to hidden neural states. This demonstrates the important confounding role of hemodynamic variability in functional networks estimated directly from fMRI time series. 4. Connectivity estimated at the neuronal level by DCM and by Granger causality after appropriate deconvolution of hemodynamics confounds indicated the driving role of S1BF cortex, as may be concluded by comparing the model evidences, at the group level, of the different classes of models tested. This finding was remarkably consistent between animals. 5. Analysis of the averaged spike-and-wave complex from depth electrode recordings indicated that the peak of the first spike in S1BF preceded by 5.5 ms and 10 ms those measured in the thalamus and the striatum, respectively. In addition, the average spike recorded in S1BF was sharper and did not show large slow wave as is the case in the thalamus and in the striatum. These characteristics indicate specific electrical signature in S1BF, potentially related to its role as a neural driver (David et al., 2008) and in agreement with electrophysiological data (see above). Our data are in agreement with other studies that used fMRI to further characterize the neural circuits involved in the generation of absence seizures in the WAG/Rij (Tenney et al., 2004) (Nersesyan et al., 2004) and in human patients with absence




epilepsy (Aghakhani, 2004) (Bai et al., 2010). Our study was the first electrophysiological validation of fMRI connectivity analyses based on Granger causality and Dynamic Causal Modelling using a well characterized animal model of functional coupling. Importantly, we demonstrated that with proper data analyses based on local deconvolution of hemodynamics, it is still possible to infer S1 as the driver of SWDs from fMRI time series (David et al., 2008; David, 2011). fMRI/EEG associated with causality analyses is thus of great help in understanding the organization of the different circuits involved in the generation, propagation and control of SWD. It should bring further insights in epileptic mechanisms when it will be possible to disentangle neuronal processes from fMRI data by using detailed generative model of fMRI signals (Aubert and Costalat, 2002; Friston et al., 2003). 4.2. Other possible uses of the GAERS model to better understand the pathophysiology of absence epilepsy Many other techniques can be applied to the GAERS model either in adults or in rat pups whether they are kept freely moving or immobilized as for intracellular electrophysiological recording or fMRI. For example, we have recently developed a preparation allowing concomitant LFP recordings and two-photon calcium imaging to further explore the S1BF cortex during epileptogenesis (Jarre et al., in prep). Similarly, LFP recording of SWDs can be easily performed in freely moving GAERS along with microdialysis (Cavarec et al., in prep) and will likely be used with optogenetic modulation (Paz et al., 2012). 5. How GAERS can help developing innovative therapeutic strategies The excellent pharmacological predictivity of the GAERS model as well as the great stability of SWDs over periods of several months make it a model of choice to develop innovative therapies for the epilepsies. Obviously, GAERS in this context is not used to develop treatment for absence epilepsy per se as this form of epilepsy remains relatively benign compared to other syndroms and because absence seizures are well controlled by several antiepileptic drugs (see above). However, the identification of a cortical “focus” generating spontaneous seizure offers a unique opportunity to perform proofs of concept for innovative strategies. Here, we described two recent sets of studies investigating new therapeutical strategies where (i) responsive deep brain stimulation was tested to interrupt SWDs and (ii) radiosurgery performed by synchrotron-generated irradiation was applied to S1 cortex as well as motor cortex and ventromedial thalamus to suppress SWD. 5.1. Deep brain stimulations High-frequency stimulation of a number of deep brain structures has been proposed during the last ten years as a way to control epileptic seizures when other therapies (AED, surgical resection, vagal nerve stimulation) are not possible (Saillet et al., 2009; Kahane and Depaulis, 2010). Deep brain stimulation (DBS) assumes that small deep nuclei can be easily modulated with a diffuse effect on cortical excitability (Theodore, 2004) and could thus constitute a better option than direct cortical stimulation of extended epileptogenic foci (Kinoshita et al., 2005). Clinical trials of DBS in epilepsy using continuous or intermittent stimulation protocols have obtained encouraging results (Loddenkemper et al., 2001; Fisher et al., 2010). Based on our previous studies on the role of the basal ganglia in the control of epileptic seizures (Depaulis et al., 1994) (Deransart and Depaulis, 2002), we first showed that high-frequency stimulation of the sub-thalamic nucleus suppresses

SWDs in GAERS (Vercueil et al., 1998). We then better characterized the optimal target and stimulation parameters necessary to interrupt SWDs and found that the substantia nigra reticulata was a better target and that 60-Hz stimulations were the most effective (Feddersen et al., 2007). However, these studies also revealed that continuous stimulation of these structures rapidly lead to a lack of response, probably due to a refractory period after each stimulation. Responsive approaches where stimulation is triggered at seizure onset after being detected online by processing brain signals (Kossoff et al., 2004; Osorio et al., 2005; Nelson et al., 2011) therefore appeared more appropriate, especially when seizure are frequent as it is the case in GAERS, as well as in some patients with absence epilepsy. Responsive stimulation offers three main potential advantages for clinical use when compared to continuous DBS: (i) a reduced power consumption of brain stimulators; (ii) a reduced neural adaptation to external stimulation; (iii) reduced behavioral or physiological side-effects (Fig. 8). We therefore examined the anti-epileptic efficacy of automated responsive DBS (rDBS) of the substantia nigra reticulata in the freely moving GAERS in which LFP were recorded for 1 h up to 24 h (Saillet et al., 2013). This required to implement an online SWDs detection algorithm based on permutation entropy (Li et al., 2007) that was coupled to an integrated electronic device allowing multi-site recording and programmable stimulations (Guillemaud et al., 2009). In this study, we obtained 100% of sensibility of the seizure detection algorithm thanks to the robust reproducibility of SWD. For short-term recordings, specificity of seizure detection was excellent (up to 99%) whereas for long-term recordings, it dropped to 72%, probably because of the increased presence of LFP components that resembled SWDs (e.g. sleep spindles and grooming artifacts). Because we privileged 100% of detection sensibility, we accepted a small decrease of specificity during the long-term recordings and did not try to optimize further the seizure detection algorithm. On the other hand, the high recurrence of SWDs in GAERS can also be perceived as a caveat regarding its predictive value on DBS protocols because the vast majority of epilepsies show much less frequent seizures. In our case, it turned out to be rather advantageous because we could evaluate habituation effects and refractory period of brain responses to rDBS (Saillet et al., 2013). During 24-h sessions, we observed a global decrease of the anti-epileptic efficacy with only 72% of interrupted SWDs, as compared to 97% during shorter sessions. During these long-lasting recordings, the rats displayed various kinds of activity (feeding, grooming, exploring) as well as periods during which SWDs occurred more frequently (e.g. in clusters) that were not clearly observed during shorter recordings. However, despite these possible experimental bias, the anti-epileptic efficacy of this approach was particularly stable over time (Saillet et al., 2013). This study provided a methodological approach for experiments with longer lasting sessions, necessary to validate the stability of rDBS over several days. It is also important to clinically address the issue of the total number of seizures recorded under responsive stimulation, as compared to OFF stimulation. Other brain targets should also be considered to further validate rDBS in epilepsy. The current development of new programmable implantable devices should make more and more feasible to assess chronic effects of rDBS in epilepsy and the GAERS model may be a valuable preclinical tool to implement these protocols that could be then applied to models of drug-resistant epilepsies. 5.2. Radiosurgery using synchrotron-generated microbeams Radiosurgery represents one of the most promising therapeutical alternative to resective surgery to cure drug-resistant epilepsies. Indeed, modern radiosurgical devices such as Gamma Knife® and Cyberknife® (Régis et al., 2006) have provided encouraging results




13

Fig. 8. Responsive deep brain stimulation. Responsive SNr stimulation was implemented using the BioMEA system (Guillemaud et al., 2009). Activity in S1 was continuously monitored and quantified using permutation entropy. On the passing of a threshold when SWDs occurred, stimulation of the substantia nigra reticulata was automatically triggered. Signal analysis was performed by means of a command software with a Matlab interfacing to command the BioMEA digital electronics. A significant reduction of SWDs was demonstrated using this strategy of control (Saillet et al., 2013).

in different types of epilepsies (Romanelli and Anschel, 2006). However, radiosurgical treatment of epileptic foci is limited by the radiosensitivity of normal tissues surrounding the lesion (Sims et al., 1999). By contrast, the low-energy photons generated by 3rd generation synchrotron sources allows the deposition of extremely steep lateral dose fall-off and provides sufficient precision to perform experimental studies in animal models of epilepsy (Studer et al., 2015). Recently, teams working at the European Synchrotron Radiation Facility (ESRF) in Grenoble used spatial fractionation of ionizing radiation in the microscopic range. The availability of microplanar 50–150 keV synchrotron-generated X-ray beams (microbeams) producing sharply defined beam edges deep in the tissues and with a high dose rate has led to investigate the effects of arrays of 25–75 ␮m-wide beam slices spaced 50–200 ␮m on centre (Slatkin et al., 1995). It soon appeared that animals surprisingly tolerate doses of hundreds, or even thousands of gray of such microbeams delivered to their brains (Slatkin et al., 1995), and tissue damage is confined to the microbeams tracks (Laissue et al., 2001). By rotating the brain around a center of rotation using a goniometer, it is possible to precisely “interlace” microbeams to deposit a high homogeneous radiation dose into discrete brain regions, with no extensions on neighboring tissue. This allows a lateral dose fall-off about 200 times greater than conventional radiotherapy (Prayson and Yoder, 2007). Its submillimetric precision of targeting, combined with the preservation of surrounding tissues, makes this procedure clinically attractive for all pathologies that require circumscribed destruction, inactivation and/or disconnection of small brain regions, even close to eloquent or vital structures. Its efficiency was first verified in rodent brain tumor

treatment and it rapidly appeared also relevant for non-cancerous brain diseases such as epilepsies. Because one of the prerequisites was to target an epileptic tissue with well-identified ictogenic neurons (see above) and presenting no gross cell loss and/or sclerosis to avoid confounding factors, the GAERS appeared as a model of choice although it is obvious that such therapy will never be applied to children with absence epilepsy. We targeted bilaterally the somatosensory and motor cortices as well as the ventrobasal thalamic nuclei. To precisely irradiate the brain volume of each of these targets, we positioned ketamine-anaesthetized GAERS rats on a Plexiglas stereotactic frame mounted on micrometric motors, allowing movements and rotations in the 3 dimensions. The bregma, at the junction of the frontal and parietal bones, can be clearly visualized on the X-ray reconstructed image of the rat’s head and used to define the coordinates of irradiation fields (Siegbahn et al., 2009). Then, the beam was spatially fractionated into an array of parallel microbeams (50 ␮m wide, 200 ␮m on-center distance) by a multislit collimator. Each side of the targeted brain regions was irradiated using 4 ports separated by a 45◦ angle and a 50-␮m step, to generate a solid dose deposition at the interlaced region. The nominal dose deposited within the target was fixed at 200 Gy following simulations (Pouyatos et al., 2013). In each animal, we verified the correct targeting by T1 -weighed MRI within the two weeks that followed irradiation. This confirmed the accuracy of radiation targeting and the minimal changes in surrounding tissues. Similarly, histopathological controls using Nissl stains and stains for myelin sheaths revealed minor structural tissue disruptions in the short-term post-irradiation phase (2 months). Outside the targets, the microbeam tracks were visible in the striatum and




Fig. 9. Effect of bilateral synchrotron-generated microbeam irradiation applied on the cortex in GAERS rats. (A) Contrast agent enhanced T1 MR image acquired 14 days after the bilateral irradiation of the somatosensory cortex (S1). Irradiated volume geometry is delimited in red. Brain structures (Mo: Motor cortex; S1: primary somatosensory cortex; CPu: Caudate Putamen nuclei) are superimposed to the MR image. (B) Nissl and myelin staining of irradiated and non-irradiated cortical tissue two months posttreatment and (C) the corresponding local field potential (LFP) recordings and time/frequency maps. (D) and (E) Electrophysiological characterization of control and irradiated neurons from the deep layers of S1 (red dots on MR image) simultaneously recorded with the surface EEG, between seizures (D) and during paroxysmal oscillations (E). Irradiated neurons displayed a severe hyperpolarization together with a reduction in action potential frequency (D) and loss of synchronized synaptic activities during seizures (E). Adapted by B. Pouyatos from Pouyatos et al. (2013). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the S1 cortex on histological brain sections, as thin clear stripes alternating with darker, nominally non-irradiated tissue swathes ≈200 ␮m spaced apart. Radiation effects were also confirmed on immuno-labeled slices with different markers of neurons, astrocytes, basal lamina, angiogenesis and glucose transporter. Although eloquent cortex was targeted bilaterally, no significant behavioral alterations were measured by Rotarod® and open-field tests 2 weeks post-irradiation. Altogether these controls confirmed that such synchrotron-generated microbeams have limited deleterious effects on brain tissue and motor or cognitive behaviors(Pouyatos et al., 2013) (Fig. 9). During the 4 months that followed irradiation, we monitored absence seizures weekly, in freely-moving GAERS, by recording LFP from depth electrodes implanted in the three target regions. These recordings showed that microbeam irradiation suppressed SWDs when the somatosensory cortex was irradiated. While non-irradiated cortical and thalamic regions could still produce epileptic discharges, the irradiated volume became isolated from the pathological activity. On the contrary irradiation of the motor cortex or ventrobasal thalamus did not significantly suppress SWDs (Pouyatos et al., 2013). We further investigated the mechanisms of the functional exclusion of the somatosensory cortex by individually recording irradiated neurons two months after irradiation using in vivo intracellular electrophysiological recordings, as described above. This approach revealed that the irradiated pyramidal neurons were strongly hyperpolarized, displayed a decreased excitability and a reduction of spontaneous synaptic activities (Pouyatos et al., 2013). These functional alterations may explain the suppression of paroxysmal oscillations within irradiated cortical

networks. This pioneering work provided the first post-irradiation electrophysiological recordings of individual neurons. As such, these recordings are a preliminary and critical step towards understanding how ionizing radiation energy delivery impacts neuronal physiology and pathophysiology. Indeed, our data should provide clues on how radiation dose should be delivered for targeting epileptic neuronal networks. 5.3. Other possible uses of the GAERS model to develop innovative therapies Because of all the characteristics indicated above, the GAERS model, as well as the WAG/Rij, offer a unique opportunity to develop innovative therapies with no or limited confounding factors like lesions, pharmacological manipulations or other biases. In particular, it is a model of choice to address long-lasting, eventually curative, treatments whether they are applied during adulthood, once the SWDs are established or during brain development, the period of epileptogenesis in this model. Gene interference during this period might be an interesting approach to both understand epileptogenesis and develop curative strategies. 6. General conclusions The GAERS model, in addition to its well recognized predictivity, has offered many methodological advantages that have allowed the use of sophisticated techniques, applied alone or in combination with others. In the next few years other new techniques will also be applied to this model, therefore increasing our knowledge




on the pathophysiology of absence epilepsy and, more generally, on the spatio-temporal and functional features of the neural circuits that generate SWD. This is quite unique in modeling CNS disease as a very same model allows to address very different questions at multiple scales and at increasing levels of complexity, from gene to long-distance networks. In addition to EEG/LFP recordings, several methods that are already used in the clinic can also be applied in this model (MRI, fMRI, DBS) therefore allowing the design of protocols that are close to clinical ones. This will certainly contribute in the future to the development of new therapies able to address epileptogenesis in addition to ictogenic processes. Acknowledgements We are greatful to our colleagues Benoit Pouyatos, Florian Studer, Colin Deransart, Isabelle Guillemain, Guillaume Jarre, Raphel Serduc, François Estève, Séverine Mahon, Pierre-Olivier Polack, Mathilde Chipaux, Mario Chavez and Michel Le van Quyen for their great contribution to the works reported in the present article. This work was supported by Inserm, ANR grants “GliEpi” and “Epirad” and Investissements d’avenir ANR-10-IAIHU-06. References Aghakhani Y. fMRI activation during spike and wave discharges in idiopathic generalized epilepsy. Brain 2004;127(Feb (5)):1127–44. Akman O, Demiralp T, Ates N, Onat FY. Electroencephalographic differences between WAG/Rij and GAERS rat models of absence epilepsy. Epilepsy Res 2010;18(Jan):1–9. Altwegg-Boussac T, Chavez M, Mahon S, Charpier S. Excitability and responsiveness of rat barrel cortex neurons in the presence and absence of spontaneous synaptic activity in vivo. J Physiol 2014;592(May (16)):3577–95. Aubert A, Costalat R. A model of the coupling between brain electrical activity, metabolism, and hemodynamics: application to the interpretation of functional neuroimaging. NeuroImage 2002;17(Nov (3)):1162–81. Avanzini G, Franceschetti S, Mantegazza M. Epileptogenic channelopathies: experimental models of human pathologies. Epilepsia 2007;48:51–64. Bai X, Vestal M, Berman R, Negishi M, Spann M, Vega C, et al. Dynamic time course of typical childhood absence seizures: EEG behavior, and functional magnetic resonance imaging. J Neurosci 2010;30(Apr (17)):5884–93. Bouilleret V, Hogan RE, Velakoulis D, Salzberg MR, Wang L, Egan GF, et al. Morphometric abnormalities and hyperanxiety in genetically epileptic rats: A model of psychiatric comorbidity? NeuroImage 2008;22(Dec):1–33. Bragin A, Engel JJ, Wilson CL, Fried I, Buzsaki G. High-frequency oscillations in human brain. Hippocampus 1999;9(2):137–42. Bruno RM, Sakmann B. Cortex is driven by weak but synchronously active thalamocortical synapses. Science 2006;312(Jun (5780)):1622–7. Caplan R, Siddarth P, Stahl L, Lanphier E, Vona P, Gurbani S, et al. Childhood absence epilepsy: behavioral, cognitive, and linguistic comorbidities. Epilepsia 2008;49(11):1838–46. Carney PW, Masterton RAJ, Flanagan D, Berkovic SF, Jackson GD. The frontal lobe in absence epilepsy: EEG-fMRI findings. Neurology 2012;78(Apr (15)):1157–65. Charpier S, Leresche N, Deniau JM, Mahon S, Hughes SW, Crunelli V. On the putative contribution of GABA(B) receptors to the electrical events occurring during spontaneous spike and wave discharges. Neuropharmacology 1999;38(11):1699–706. Chipaux M, Charpier S, Polack PO. Chloride-mediated inhibition of the ictogenic neurones initiating genetically-determined absence seizures. Neuroscience 2011;Sep (192):642–51. Chipaux M, Vercueil L, Kaminska A, Mahon S, Charpier S. Persistence of cortical sensory processing during absence seizures in human and an animal model: evidence from EEG and intracellular recordings. PLoS ONE 2013;8(3):e58180. Constantinople CM, Bruno RM. Effects mechanisms of wakefulness on local cortical networks. Neuron 2011;69(Mar (6)):1061–8. Contreras D, Destexhe A, Steriade M. Intracellular and computational characterization of the intracortical inhibitory control of synchronized thalamic inputs in vivo. J Neurophysiol 1997;78(1):335–50. ˝ Cope DW, Di Giovanni G, Fyson SJ, Orbán G, Errington AC, Lorincz ML, et al. Enhanced tonic GABAA inhibition in typical absence epilepsy. Nat Med 2009;22(Nov):1–8. Crunelli V, Leresche N. Childhood absence epilepsy: genes, channels, neurons and networks. Nat Rev Neurosci 2002;3(5):371–82. David O. fMRI connectivity, meaning and empiricism. NeuroImage 2011;58(Sep (2)):306–9. David O, Cosmelli D, Friston KJ. Evaluation of different measures of functional connectivity using a neural mass model. NeuroImage 2004;21(2):659–73. David O, Guillemain I, Saillet S, Reyt S, Deransart C, Segebarth C, et al. Identifying neural drivers with functional MRI: an electrophysiological validation. PLoS Biol 2008;6(12):e315–2697.

15

de Munck JC, Vijn PC, Lopes da Silva FH. A random dipole model for spontaneous brain activity. IEEE Trans Biomed Eng 1992;39(8):791–804. Depaulis A, Hamelin S. Animal models for mesiotemporal lobe epilepsy: the end of a misunderstanding? Rev Neurol 2015;171(3):217–26. Depaulis A, van Luijtelaar G. Genetic models of absence epilepsy in the rat. In: Pitkänen A, Schwartzkroin P, Moshe S, editors. Models of seizures and epilepsy. Amsterdam, Oxford: Elsevier Academic; 2005. p. 233–48. Depaulis A, Vergnes M, Marescaux C. Endogenous control of epilepsy: the nigral inhibitory system. Prog Neurobiol 1994;42(Jul (1)):33–52. Deransart C, Depaulis A. The control of seizures by the basal ganglia? A review of experimental data. Epileptic Disord 2002;4(Dec (Suppl. 3)):S61–72. Deransart C, Vercueil L, Marescaux C, Depaulis A. The role of basal ganglia in the control of generalized absence seizures. Epilepsy Res 1998;32(Sep (1–2)): 213–23. Destexhe A, Rudolph M, Paré D. The high-conductance state of neocortical neurons in vivo. Nat Rev Neurosci 2003;4(9):739–51. Eccles JC. The physiology of synapses. Berlin: Springer; 1964. Feddersen B, Vercueil L, Noachtar S, David O, Depaulis A, Deransart C. Controlling seizures is not controlling epilepsy: a parametric study of deep brain stimulation for epilepsy. Neurobiol Dis 2007;27(Sep (3)):292–300. Feldmeyer D, Brecht M, Helmchen F, Petersen CCH, Poulet JFA, Staiger JF, et al. Barrel cortex function. Prog Neurobiol 2013;103(Apr):3–27. Fisher R, Salanova V, Witt T, Worth R, Henry T, Gross R, et al. Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory epilepsy. Epilepsia 2010;51(Mar (5)):899–908. Flacknell P. Laboratory animal anaesthesia. London: Academic Press; 1996. Frankel WN. Genetics of complex neurological disease: challenges and opportunities for modeling epilepsy in mice and rats. Trends Genet 2009;25(8):361–7. Friston KJ, Harrison L, Penny W. Dynamic causal modelling. NeuroImage 2003;19(Aug (4)):1273–302. Girod S, Chabrol T, Pouyatos B, Depaulis A, Guillemain I. Early development of seizures in the somatosensory cortex in a genetic model of absence epilepsy in the rat: ontogenesis revisited. In: 27th Annual meeting of the society for neuroscience; 2012. Glauser TA, Cnaan A, Shinnar S, Hirtz DG, Dlugos D, Masur D, et al. Ethosuximide, valproic acid, and lamotrigine in childhood absence epilepsy. N Engl J Med 2010;362(Mar (9)):790–9. Goebel R, Roebroeck A, Kim D-S, Formisano E. Investigating directed cortical interactions in time-resolved fMRI data using vector autoregressive modeling and Granger causality mapping. Magn Reson Imaging 2003;21(Dec (10)):1251–61. Gorji A, Speckmann E-J. Epileptiform EEG spikes and their functional significance. Clin EEG Neurosci 2009;40(Oct (4)):230–3. Guillemain I, Kahane P, Depaulis A. Animal models to study aetiopathology of epilepsy: what are the features to model? Epileptic Disord 2012;14(Sep):217–25. Guillemaud R, Bêche JF, Billoint O, Bonnet S, Gharbi S, Rostaing JP, et al. A multichannel platform for recording and stimulation of large neuronal structures. IRBM 2009;30(Nov (5–6)):226–33. Hamandi K, Laufs H, Nöth U, Carmichael DW, Duncan JS, Lemieux L. BOLD and perfusion changes during epileptic generalised spike wave activity. NeuroImage 2008;39(Jan (2)):608–18. He C, Chen F, Li B, Hu Z. Neurophysiology of HCN channels: from cellular functions to multiple regulations. Prog Neurobiol 2014;112:1–23. Holmes MD, Brown M, Tucker DM. Are “generalized” seizures truly generalized? Evidence of localized mesial frontal and frontopolar discharges in absence. Epilepsia 2004;45(12):1568–79. Hughes JR. Progress in predicting seizure episodes with nonlinear methods. Epilepsy Behav 2008;12:128–35. Iadecola C, Nedergaard M. Glial regulation of the cerebral microvasculature. Nat Neurosci 2007;10(11):1369–76. Inoue M, Ates N, Vossen JM, Coenen AM. Effects of the neuroleptanalgesic fentanylfluanisone (Hypnorm) on spike-wave discharges in epileptic rats. Pharmacol Biochem Behav 1994;48(Jun (2)):547–51. Jones NC, Salzberg MR, Kumar G, Couper A, Morris MJ, O’Brien TJ. Elevated anxiety and depressive-like behavior in a rat model of genetic generalized epilepsy suggesting common causation. Exp Neurol 2008;209(1):254–60. Kahane P, Depaulis A. Deep brain stimulation in epilepsy: what is next? Curr Opin Neurol 2010;23(2):177–82. Kinoshita M, Ikeda A, Matsuhashi M, Matsumoto R, Hitomi T, Begum T, et al. Electric cortical stimulation suppresses epileptic and background activities in neocortical epilepsy and mesial temporal lobe epilepsy. Clin Neurophysiol 2005;116(6):1291–9. Klein J. Dysregulation of sodium channel expression in cortical neurons in a rodent model of absence epilepsy. Brain Res 2004;1000(Mar (1–2)):102–9. Kossoff EH, Ritzl EK, Politsky JM, Murro AM, Smith JR, Duckrow RB, et al. Effect of an external responsive neurostimulator on seizures and electrographic discharges during subdural electrode monitoring. Epilepsia 2004;45(12):1560–7. Kuisle M, Wanaverbecq N, Brewster AL, Frère SGA, Pinault D, Baram TZ, et al. Functional stabilization of weakened thalamic pacemaker channel regulation in rat absence epilepsy. J Physiol 2006;575(Aug (1)):83–100. Laissue J, Blattmann H, Di Michiel M, Slatkin DN, Lyubimova N, Guzman R, et al. The weanling piglet cerebellum: a surrogate for tolerance to MRT (microbeam radiation therapy) in pediatric neuro-oncology. Proc SPIE 2001;4508(Jul):65–73. Langlois M, Polack PO, Bernard H, David O, Charpier S, Depaulis A, et al. Involvement of the thalamic parafascicular nucleus in mesial temporal lobe epilepsy. J Neurosci 2010;30(Dec (49)):16523–35.




Lannes B, Micheletti G, Vergnes M, Marescaux C, Depaulis A, Warter JM. Relationship between spike-wave discharges and vigilance levels in rats with spontaneous petit mal-like epilepsy. Neurosci Lett 1988;94:187–91. Li X, Ouyang G, Richards DA. Predictability analysis of absence seizures with permutation entropy. Epilepsy Res 2007;77(Oct (1)):70–4. Loddenkemper T, Pan A, Neme S, Baker KB, Rezai AR, Dinner DS, et al. Deep brain stimulation in epilepsy. J Clin Neurophysiol 2001;18(6):514–32. Loiseau P, Duché B, Pédespan JM. Absence epilepsies. Epilepsia 1995;36(12):1182–6. Lopes da Silva F, Pijn JP, Boeijinga P. Interdependence of EEG signals: linear vs. nonlinear associations and the significance of time delays and phase shifts. Brain Topogr 1989;2(1–2):9–18. Mahon S, Deniau JM, Charpier S. Relationship between EEG potentials and intracellular activity of striatal and cortico-striatal neurons: an in vivo study under different anesthetics. Cereb Cortex 2001;11(4):360–73. Manning J-PA, Richards DA, Leresche N, Crunelli V, Bowery NG. Cortical-area specific block of genetically determined absence seizures by ethosuximide. Neuroscience 2004;123(1):5–9. Marescaux C, Micheletti G, Vergnes M, Depaulis A, Rumbach L, Warter JM. A model of chronic spontaneous petit mal-like seizures in the rat: comparison with pentylenetetrazol-induced seizures. Epilepsia 1984;25(Jun (3)):326–31. Marescaux C, Vergnes M, Depaulis A. Genetic absence epilepsy in rats from Strasbourg—a review. J Neural Transm—Suppl 1992a;35:37–69. Marescaux C, Vergnes M, Liu Z, Depaulis A, Bernasconi R. GABAB receptor involvement in the control of genetic absence seizures in rats. Epilepsy Res Suppl 1992b;9:131–8. Meeren H, van Luijtelaar G, Lopes da Silva F, Coenen A. Evolving concepts on the pathophysiology of absence seizures: the cortical focus theory. Arch Neurol 2005;62(3):371–6. Meeren HKM, Pijn JPM, Van Luijtelaar ELJM, Coenen AML, Lopes da Silva FH. Cortical focus drives widespread corticothalamic networks during spontaneous absence seizures in rats. J Neurosci 2002;22(Feb (4)):1480–95. 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(Aug (10)):2000–10. Nelson TS, Suhr CL, Lai A, Halliday AJ, Freestone DR, McLean KJ, et al. Exploring the tolerability of spatiotemporally complex electrical stimulation paradigms. Epilepsy Res 2011;25(Jul):1–9. Nersesyan H, Hyder F, Rothman DL, Blumenfeld H. Dynamic fMRI and EEG recordings during spike-wave seizures and generalized tonic–clonic seizures in WAG/Rij rats. J Cereb Blood Flow Metab 2004;24(6):589–99. Noebels JL. Single-gene models of epilepsy. Adv Neurol 1999;79:227–38. Osorio I, Frei MG, Sunderam S, Giftakis J, Bhavaraju NC, Schaffner SF, et al. Automated seizure abatement in humans using electrical stimulation. Ann Neurol 2005;57(2):258–68. Panayiotopoulos C. Typical absence seizures and their treatment. Arch Dis Childhood 1999;81(4):351–5. Parker APJ, Agathonikou A, Robinson RO, Panayiotopoulos CP. Inappropriate use of carbamazepine and vigabatrin in typical absence seizures. Dev Med Child Neurol 1998;40(8):517–9. Paz JT, Chavez M, Saillet S, Deniau J-M, Charpier S. Activity of ventral medial thalamic neurons during absence seizures and modulation of cortical paroxysms by the nigrothalamic pathway. J Neurosci 2007;27(Jan (4)):929–41. Paz JT, Davidson TJ, Frechette ES, Delord B, Parada I, Peng K, et al. Closed-loop optogenetic control of thalamus as a tool for interrupting seizures after cortical injury. Nat Neurosci 2012;16(Nov (1)):64–70. Pinault D. A novel single-cell staining procedure performed in vivo under electrophysiological control: morpho-functional features of juxtacellularly labeled thalamic cells and other central neurons with biocytin or Neurobiotin. J Neurosci Methods 1996;65(Apr (2)):113–36. Pinault D, Leresche N, Charpier S, Deniau JM, Marescaux C, Vergnes M, et al. Intracellular recordings in thalamic neurones during spontaneous spike and wave discharges in rats with absence epilepsy. J Physiol—London 1998;509:449–56. Polack P-O, Charpier S. Ethosuximide converts ictogenic neurons initiating absence seizures into normal neurons in a genetic model. Epilepsia 2009;50(7):1816–20. Polack PO, Guillemain I, Hu E, Deransart C, Depaulis A, Charpier S. Deep layer somatosensory cortical neurons initiate spike-and-wave discharges in a genetic model of absence seizures. J Neurosci 2007;27(Jun (24)):6590–9. Polack PO, Mahon S, Chavez M, Charpier S. Inactivation of the somatosensory cortex prevents paroxysmal oscillations in cortical and related thalamic neurons in a genetic model of absence epilepsy. Cereb Cortex 2009;19(Aug (9)):2078–91. Porter RJ. The absence epilepsies. Epilepsia 1993;34:s42–8. Pouyatos B, Serduc R, Chipaux M, Chabrol T, Bräuer-Krisch E, Nemoz C, et al. Synchrotron X-ray interlaced microbeams suppress paroxysmal oscillations in neuronal networks initiating generalized epilepsy. Neurobiol Dis 2013;51(C):152–60. Powell KL, Cain SM, Ng C, Sirdesai S, David LS, Kyi M, et al. A Cav3.2 T-Type calcium channel point mutation has splice-variant-specific effects on function and segregates with seizure expression in a polygenic rat model of absence epilepsy. J Neurosci 2009;29(Jan (2)):371–80. Powell KL, Tang H, Ng C, Guillemain I, Dieuset G, Dezsi G, et al. Seizure expression, behavior, and brain morphology differences in colonies of Genetic Absence Epilepsy Rats from Strasbourg. Epilepsia 2014;55(Nov (12)):1959–68. Prayson RA, Yoder BJ. Clinicopathologic findings in mesial temporal sclerosis treated with gamma knife radiotherapy. Ann Diagnostic Pathol 2007;11(Feb (1)):22–6.

Quian Quiroga R, Kraskov A, Kreuz T, Grassberger P. Performance of different synchronization measures in real data: a case study on electroencephalographic signals. Phys Rev E: Stat Nonlin Soft Matter Phys 2002;65(4 (Pt 1)):0419–423. Rasch M, Logothetis NK, Kreiman G. From neurons to circuits: linear estimation of local field potentials. J Neurosci 2009;29(Nov (44)), 1378-5-96. Régis J, Scavarda D, Tamura M, Nagayi M, Villeneuve N, Bartolomei F, et al. Epilepsy related to hypothalamic hamartomas: surgical management with special reference to gamma knife surgery. Childs Nerv Syst 2006;22(Jun (8)): 881–95. Romanelli P, Anschel DJ. Radiosurgery for epilepsy. Lancet Neurol 2006;5(7): 613–20. Rudolf G, Bihoreau MT, Godfrey RF, Wilder SP, Cox RD, Lathrop M, et al. Polygenic control of idiopathic generalized epilepsy phenotypes in the genetic absence rats from Strasbourg (GAERS). Epilepsia 2004;45(4):301–8. Sadleir LG, Farrell K, Smith S, Connolly MB, Scheffer IE. Electroclinical features of absence seizures in childhood absence epilepsy. Neurology 2006;67(Aug (3)):413–8. Sahin M, Bowen WD, Donoghue JP. Location of nicotinic and muscarinic cholinergic and mu-opiate receptors in rat cerebral neocortex: evidence from thalamic and cortical lesions. Brain Res 1992;579(May (1)):135–47. Saillet S, Langlois M, Feddersen B, Minotti L, Vercueil L, Chabardès S, et al. Manipulating the epileptic brain using stimulation: a review of experimental and clinical studies. Epileptic Disord 2009;11(Jun (2)):100–12. Saillet S, Gharbi S, Charvet G, Deransart C, Guillemaud R, Depaulis A, et al. Neural adaptation to responsive stimulation: a comparison of auditory and deep brain stimulation in a rat model of absence epilepsy. Brain Stimul 2013;6(May (3)):241–7. Seidenbecher T, Pape HC. Contribution of intralaminar thalamic nuclei to spike-andwave-discharges during spontaneous seizures in a genetic rat model of absence epilepsy. Eur J Neurosci 2001;13(Apr (8)):1537–46. Siegbahn EA, Bräuer-Krisch E, Bravin A, Nettelbeck H, Lerch ML, Rosenfeld AB. MOSFET dosimetry with high spatial resolution in intense synchrotron-generated X-ray microbeams. Med Phys 2009;36(Mar):1128–37. Simonato M, Brooks-Kayal A, Engel J Jr, Aristea G, Jensen FE, Moshé SL, et al. Personal view the challenge and promise of anti-epileptic therapy development in animal models. Lancet Neurol 2014;13(Aug (9)):949–60. Simons DJ, Carvell GE. Thalamocortical response transformation in the rat vibrissa/barrel system. J Neurophysiol 1989;61(2):311–30. Simons DJ, Carvell GE, Kyriazi HT, Bruno RM. Thalamocortical conduction times and stimulus-evoked responses in the rat whisker-to-barrel system. J Neurophysiol 2007;98(Sep (5)):2842–7. Sims E, Doughty D, Macaulay E, Royle N, Wraith C, Darlison R, et al. Stereotactically delivered cranial radiation therapy: a ten-year experience of linac-based radiosurgery in the UK. Clinical Oncology 1999;11(5):303–20. Slaght SJ. On the activity of the corticostriatal networks during spike-and-wave discharges in a genetic model of absence epilepsy. J Neurosci 2004;24(Jul (30)):6816–25. Slaght SJ, Leresche N, Deniau J-M, Crunelli V, Charpier S. Activity of thalamic reticular neurons during spontaneous genetically determined spike and wave discharges. J Neurosci 2002;22(Mar (6)):2323–34. Slatkin DN, Spanne P, Dilmanian FA, Gebbers JO, Laissue JA. Subacute neuropathological effects of microplanar beams of X-rays from a synchrotron wiggler. Proc Natl Acad Sci USA 1995;92(Sep (19)):8783–7. Studer F, Serduc R, Pouyatos B, Chabrol T, Bräuer-Krisch E, Donzelli M, et al. Synchrotron X-ray microbeams: a promising tool for drug-resistant epilepsy treatment. Phys Med 2015(Apr). Tenney JR, Duong TQ, King JA, Ferris CF. FMRI of brain activation in a genetic rat model of absence seizures. Epilepsia 2004;45(6):576–82. Theodore W. Brain stimulation for epilepsy. Lancet Neurol 2004;3(2):111–8. Timofeev I, Steriade M. Neocortical seizures: initiation, development and cessation. Neuroscience 2004;123(Jan (2)):299–336. Tsakiridou E, Bertollini L, deCurtis M, Avanzini G, Pape HC. Selective increase in T-Type calcium conductance of reticular thalamic neurons in a rat model of absence epilepsy. J Neurosci 1995;15(4):3110–7. Vercueil L, Benazzouz A, Deransart C, Bressand K, Marescaux C, Depaulis A, et al. High-frequency stimulation of the sub-thalamic nucleus suppresses absence seizures in the rat: comparison with neurotoxic lesions. Epilepsy Res 1998;31(1):39–46. Vergnes M, Marescaux C, Boehrer A, Depaulis A. Are rats with genetic absence epilepsy behaviorally impaired? Epilepsy Res 1991;9(Jul (2)):97–104. Vergnes M, Marescaux C, Depaulis A. Mapping of spontaneous spike and wave discharges in Wistar rats with genetic generalized non-convulsive epilepsy. Brain Res 1990;523(Jul (1)):87–91. Vergnes M, Marescaux C, Depaulis A, Micheletti G, Warter JM. Ontogeny of spontaneous petit mal-like seizures in Wistar rats. Brain Res 1986;395(Apr (1)):85–7. Vergnes M, Marescaux C, Micheletti G, Depaulis A, Rumbach L, Warter JM. Enhancement of spike and wave discharges by GABAmimetic drugs in rats with spontaneous petit-mal-like epilepsy. Neurosci Lett 1984;44(1):91–4. Vergnes M, Marescaux C, Micheletti G, Reis J, Depaulis A, Rumbach L, et al. Spontaneous paroxysmal electroclinical patterns in rat: a model of generalized non-convulsive epilepsy. Neurosci Lett 1982;33(Nov (1)):97–101. Yagüe JG, Cavaccini A, Errington AC, Crunelli V, Giovanni GD. Dopaminergic modulation of tonic but not phasic GABAA-receptor-mediated current in the ventrobasal thalamus of Wistar and GAERS rats. Exp Neurol 2013;247:1–7.


Genetic absence epilepsy in rats from Strasbourg--a review.

Age-dependent Electroencephalographic Differences in the Genetic Absence Epilepsy Rats from Strasbourg (GAERS) Model of Absence Epilepsy.

Evaluating whole genome sequence data from the Genetic Absence Epilepsy Rat from Strasbourg and its related non-epileptic strain.

Proteomic changes in the cortex membrane fraction of genetic absence epilepsy rats from Strasbourg.

Anxiety and locomotion in Genetic Absence Epilepsy Rats from Strasbourg (GAERS): inclusion of Wistar rats as a second control.

Are Idiopathic Generalized Epilepsies Focal?

Genetic Causes of Generalized Epilepsies.

Genetic Generalized Epilepsy.

Focal seizure symptoms in idiopathic generalized epilepsies.

Cognitive impairments and neuronal injury in different brain regions of a genetic rat model of absence epilepsy.

Identifying Corticothalamic Network Epicenters in Patients with Idiopathic Generalized Epilepsy.

Increasing blurriness of the borders between focal and generalized as well as cryptogenic and idiopathic epilepsies in defining the role for focal epilepsy surgery.

Wake System?

Social cognition dysfunctions in patients with epilepsy: Evidence from patients with temporal lobe and idiopathic generalized epilepsies.

A Network Biology Approach to Decipher Stress Response in Bacteria Using Escherichia coli As a Model.

Genetics of idiopathic generalized epilepsy: an overview.

Functional neuroimaging abnormalities in idiopathic generalized epilepsy.

Low-dose sodium valproate in the treatment of idiopathic generalized epilepsies.

Molecular Pathogenic Basis for GABRG2 Mutations Associated With a Spectrum of Epilepsy Syndromes, From Generalized Absence Epilepsy to Dravet Syndrome.

Genetic variants of synaptic vesicle and presynaptic plasma membrane proteins in idiopathic generalized epilepsy.

Comparative proteomic approach in rat model of absence epilepsy.

Evaluation of an automated spike-and-wave complex detection algorithm in the EEG from a rat model of absence epilepsy.

progenitor cell proliferation in the hippocampus are altered during the development of absence epilepsy in a genetic animal model.

Eyelid twitching seizures and generalized tonic-clonic convulsions: a syndrome of idiopathic generalized epilepsy.