Electrical stimulation of the epileptic focus in absence epileptic WAG/Rij rats: assessment of local...

Please cite this article in press as: Lüttjohann A, et al., Electrical stimulation of the epileptic focus in absence epileptic WAG/Rij rats:Assessment of local and network excitability, Neuroscience (2011), doi: 10.1016/j.neuroscience.2011.04.038

Neuroscience xx (2011) xxx

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455

ELECTRICAL STIMULATION OF THE EPILEPTIC FOCUS IN ABSENCEEPILEPTIC WAG/RIJ RATS: ASSESSMENT OF LOCAL AND NETWORK

1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253

A. LÜTTJOHANN,a S. ZHANG,b

R. DE PEIJPERa AND G. VAN LUIJTELAARa*aDonders Center for Cognition, Donders Institute for Brain, Cognitionand Behaviour, Radboud University Nijmegen, Nijmegen, The Neth-erlandsbSecond Ward, Beijing Institute of Functional Neurosurgery, XuanwuHospital, Capital Medical University, Beijing, PR China

Abstract—[Objective] The study aims to investigate whetherthere is a higher excitability in the deep cortical layers of theperi-oral region of the somatosensory cortex as compared toother cortical regions in absence epileptic WAG/Rij rats andwhether this is unique for this type of epileptic rats, as wouldbe predicted by the cortical focus theory of absence epilepsy.[Methods] Excitability of cortical structures was assessed ina double pulse paradigm (inter-pulse interval 400 ms, 400 �s

ulse duration, varying stimulation intensities (20–100 �A)).Electrical stimulation was applied to the subgranular layers ofthe somatosensory and motor cortex of freely moving WAG/Rijand control Wistar rats. Electrical evoked potentials (EEPs) andafterdischarges (ADs) were recorded during wakefulness,drowsiness and non-REM sleep. [Results] WAG/Rij rats, stimu-lated in the somatosensory cortex, showed higher amplitudesfor the N1 and N3 components of the EEPs as compared toWAG/Rij rats stimulated in the motor cortex. This effect waspresent in all states of alertness and at all tested intensities.In addition, this effect was not (N1) present or to much lessextent (N3) present in nonepileptic control rats. Stimulation-induced 8 Hz ADs were predominantly found in WAG/Rij rats.ADs were longer after stimulation in the somatosensory thanin the motor cortex and preferentially occurred during drows-iness. [Conclusion] There is a heightened excitability in thedeep layer neurons of the perioral region of somatosensorycortex, which is unique for WAG/Rij rats. Moreover, the pres-ence of 8 Hz ADs might point toward additional changes inthe cortico-thalamo-cortical network. Drowsiness is an excel-lent state for 8 Hz ADs, mimicking spike and wave discharges(SWDs). The results are in good agreement with the cortical-focus theory of absence epilepsy. © 2011 Published byElsevier Ltd on behalf of IBRO.

Key words: somatosensory cortex, subgranular layer, elec-trical stimulation, electrical evoked potentials, WAG/Rij rats,afterdischarges, sleep-wake and drowsiness.

In the past 60 years different theories on the origin of spikeand wave discharges (SWDs), the electroencephalo-graphic hallmark of absence epilepsy, have been pro-posed. These include the centrencephalic theory, propos-

*Corresponding author. Tel: �31-024-3615621.-mail address: g.vanluijtelaar@donders.ru.nl (G. van Luijtelaar).

tapraid3/zpn-nsc/zpn-nsc/zpn99907/zpn5562d07z routa

ing a subcortical generator in the intralaminar thalamicnuclei as a continuation of the brainstem reticular activa-tion system (Penfield and Jsper, 1954); the thalamic clocktheory suggesting a seizure generator in the reticular tha-lamic nucleus (RTN), which imposes its rhythm to thethalamo-cortical relay cells (Buzsáki, 1991); the corticaltheory which emphasize the role of the cortex in the gen-eration of SWDs as a result of cortical abnormalities(Lüders et al., 1984; Niedermeyer, 1972) or as a second-arily generalized discharge arising near a lesion in thefrontal cortex (Bancaud, 1969). Currently the most widelyaccepted view on the origin of SWDs is that sleep spindles,which are the result of increased activity of GABAergicRTN cells that hyperpolarize thalamocortical cells and setthem into a burst firing mode (Steriade, 2003), are modifiedby a globally, hyperexcitable cortex into pathologicalSWDs (cortico-reticular theory) (Gloor, 1968; Kostopoulos,2000; Meeren et al., 2005).

Recent data in two genetic animal models of absenceepilepsy (WAG/Rij and GAERS rats), however, questionthese ideas: signal analytical techniques, applied on ECoGsignals containing SWDs obtained from a cortical grid incombination with thalamic recordings, demonstrated thatthe start of SWDs can be observed in the perioral region ofthe somatosensory cortex and that only subsequently therest of the cortex and thalamus get involved (Meeren et al.,2002; van Luijtelaar et al., 2011). In addition, Polack et al.(2007, 2009) described in GAERS that cells in corticallayer V and VI of the somatosensory cortex are depolar-ized interictally and show an increased firing rate alreadybefore SWDs can be seen in local field potentials. Theseresults fit excellently in the cortical focus theory on ab-sence epilepsy stating that the perioral region of the so-matosensory cortex contains a “hot spot” where SWDs aregenerated (Meeren et al., 2002; van Luijtelaar and Sit-nikova, 2006). This theory implies that the excitability in thecortical focus should be higher as compared to other cor-tical regions. On top, this should be unique for absenceepileptic rats. This was tested in the current experiment invivo. Freely moving rats received electrical double pulsestimulation in the either the focal zone or the motor cortex,outside but adjacent to the proposed epileptic focus. Re-sulting electrical evoked potentials (EEPs) were recordedlocally at the somatosensory and motor cortex. Local ex-citability in the deep cortical layers was determined bymeasuring the amplitude of the early components of theseEEPs. The slower components of the evoked potentialsmight reflect cortical network or even subcortical network

S�1 5/10/11 6:09 Art: 12902

pewwldp

pEuRpwc

bAaRmf

icttpEcdrTRR

A. Lüttjohann et al. / Neuroscience xx (2011) xxx2

5657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114

5657585960616263646566676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113

Since the amplitude of auditory and visual evokedotentials is known to change with vigilance states (Co-nen, 1995; Meeren et al., 1998) and since it is not knownhether this is also the case for EEPs, measurementsere done in three different states of vigilance. This al-

owed us to investigate whether excitability may changeuring the natural vigilance states and to control for thisutative confounding factor on EEP amplitude.

Electrical stimulation of the sensorimotor cortex maye followed by different types of afterdischarges (ADs);

heir type and nature depends on stimulation intensityMares and Kubova, 2006). The amount and duration oflectrically-induced ADs is used as an index for corticalnd subcortical network excitability and recruitment.

In accordance with the cortical focus theory it is ex-ected that the amplitude of the early components ofEPs, indicating excitability, is higher in WAG/Rij rats stim-lated in the somatosensory cortex as compared to WAG/ij rats stimulated in the motor cortex. In addition, therobability of occurrence and duration of ADs will be longerhen stimulated in the somatosensory cortex versus motorortex and will depend on the vigilance state.

All measurements were done in WAG/Rij rats andon-epileptic control Wistar rats in order to test whether

hese effects are unique for rats with absence seizures.

EXPERIMENTAL PROCEDURES

Animals

Subjects, male WAG/Rij rats and Wistar rats, age 12–14 months,mean body weight of 350 g (range 324–380 g) were born andraised in the laboratory of Biological Psychology, Donders Centrefor Cognition at Radboud University Nijmegen, the Netherlands.Housing conditions were in accordance with standard laboratoryconditions (two animals per cage including cage enrichment (En-viro Dry®)). Rats were kept at 12:12 light/dark cycle (light off phaseetween 8.00 and 20.00 h) and had free access to food and water.fter surgery rats were housed individually. The experiment waspproved by the Ethical Committee on Animal Experimentation ofadboud University Nijmegen (RU-DEC). Efforts were done toinimize the number of used animals and to make the discomfort

or the animals as minimal as possible.

Stereotactic surgery

Stereotactical surgery was performed under isoflurane anesthe-sia: WAG/Rij and Wistar rats were divided in two groups. Group 1(eight WAG/Rij and eight Wistar rats): two stimulation electrodewires (anode and cathode) were implanted in the deep layers ofthe somatosensory cortex, right hemisphere (A/P: �0 mm, M/L:�5 mm, H: �4.5 mm). The tip distance of anode and cathode wasabout 1 mm in order to achieve local stimulation of the corticalfocus. Two active EEG recording electrodes were additionallyimplanted in these rats: one of these recording electrodes wasplaced near the stimulation site in the deep layers of the somato-sensory cortex (tip distance between recording and the neareststimulation electrode was 1 mm, all three wires of the tripolarelectrode set (Plastic One, Roanoke, VA, USA) were oriented inA/P direction). The second recording electrode was epidurallyimplanted on the remote motor cortex of the right hemisphere(A/P: �2 mm, M/L: �2 mm). A reference and a ground electrodewere implanted on top of the cerebellum.

Group 2 (nine WAG/Rij and eight Wistars): two stimulation

Please cite this article in press as: Lüttjohann A, et al., Electrical stimulAssessment of local and network excitability, Neuroscience (2011), doi:

electrode were implanted in the deep layers of the motor cortex,right hemisphere (A/P: �2 mm, M/L: �2 mm, H: �3 mm). Tipdistance between the recording and the nearest stimulation elec-trodes was again 1 mm; the orientation (A/P) of the three depthelectrodes was the same as in animals from group I. This elec-trode configuration allowed stimulation and recording of the con-trol area adjacent to the cortical focus. A second EEG recordingelectrode was epidurally implanted on the remote somatosensorycortex (A/P: �0 mm, M/L: �5 mm). Again a reference and aground electrode were implanted on top of the cerebellum.

All coordinates were determined according to the stereotacticatlas of Paxinos and Watson (1998). Electrode assemblies werefixed to the scull with the aid of dental cement. The rats werepreoperative injected with Atropine and Rimadyl®, and again withRimadyl® postoperatively (24 and 48 hours after surgery). Ratswere allowed to recover for 2 weeks.

EEG recording and stimulation

Rats were placed individually in a 20�35�25 cm3 Plexiglas reg-stration box and connected to the recording leads for EEG re-ording and electrical brain stimulation. The leads were attachedo a swivel-contact, which allowed registration and stimulation inhe freely moving animals. The EEG signals were amplified with ahysiological amplifier (TD 90087, Radboud University Nijmegen,lectronic Research Group), filtered by a band pass filter withut-off points at 1(HP) and 100(LP) and a 50 Hz Notch filter, andigitalized with a constant sample rate of 256 Hz by WINDAQ®-ecording-system (DATAQ-Instruments,Inc., Akron, OH, USA).he movements of the rat were registered by a Passive Infraredegistration system (PIR, RK2000DPC LuNAR PR Ceiling Mount,okonet RISCO Group S.A., Drogenbos, Belgium).

Electrical stimulation was provided with the aid of a program-able Stimulus Generator (TD 10075 Electronic Research Group,adboud University Nijmegen), and an Isolated Constant Currentnit. All rats received electrical double pulse stimulation with aulse-duration of 0.4 ms, and an interpulse interval of 400 ms.

Stimulation in WAG/Rij rats was given during wakefulness,rowsiness and non-REM sleep by a trained EEG analyst. Doubleulses of five different intensities (20, 40, 60, 80,100 �A) were

given per state of alertness resulting in a total of 15 stimulationconditions. The order of the stimulation conditions was random-ized. WAG/Rij rats received a mean of 30 (range 10–76) subse-quent double-pulse pairs during each stimulation condition with avariable inter pair interval of at least ten seconds.

All responses to stimuli of the same stimulation condition(stimuli of the same intensity and delivered during the same stateof alertness) were off line averaged resulting in 15 EEPs per rat.

Time-points of stimulation were retained by time point mark-ers which were conveyed by the stimulation unit to the WINDAQ-recording-system and displayed in the form of block pulses in anauxiliary channel in the data acquisition system. However, given thatthe signal of the time point markers was, in contrast to the EEG,unfiltered, a time delay of about 4 ms was introduced in the EEGsignal (this was empirically verified). This implies that the recordedlatencies of the resulting EEPs are delayed by 4 ms. Furthermore, forEEPs typical stimulation artifact is not visible in the recordings due tothe filters and sampling properties.

Wistar rats, which were used to determine whether a possibledifference in the excitability between somatosensory and motorcortex is a unique property of absence epileptic rats or a generalphenomenon for sensory and motor cortices, received only doublepulse stimulation of 60 �A during wakefulness.

The state of alertness was determined during recording, byobserving the animal’s behavior and EEG. The criteria for wake-fulness were a moving animal visualized in the recordings by anactive PIR and a high frequency, low amplitude EEG. For non-REM sleep the animal was immobile with a flat PIR and a low

ation of the epileptic focus in absence epileptic WAG/Rij rats:10.1016/j.neuroscience.2011.04.038

S�1 5/10/11 6:09 Art: 12902

sgrwvb2Wsf

A. Lüttjohann et al. / Neuroscience xx (2011) xxx 3

115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173

115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172

animal was motionless with a flat PIR with only minor movementsand an intermediate EEG with respect to amplitude and presenceof slow activity for at least 10 s before stimulation.

After termination of the experiment, all rats were perfused forhistological verification of the electrode locations. Brains werefixed in a 30% sucrose solution, 0.1 ml TBS, cut in 40 �m coronalslices with a microtome, and stained with Cresyl Violet.

Data analysis

BrainVision® analyzer (Brain Products GmbH, Gilching, Germany)was used for the offline averaging of EEPs. Single trial responseswere averaged per intensity and vigilance state resulting in 15EEPs per rat. A mean of 30 responses to stimulation (range10–76) were averaged to obtain one EEP. The main componentsof the EEPs were identified and categorized by their latencies. Theamplitudes of these components were statistically analyzed.

A repeated measures ANOVA with amplitude as dependentvariable, stimulation site (somatosensory and motor cortex) asbetween subjects factor and state of vigilance (wakefulness,drowsiness and sleep), intensity (20–100 �A) and stimulus (stim-lus 1, stimulus 2) as within subjects factors was used for the EEPata of WAG/Rij rats.

EEP data of the Wistar rats, which were, due to their controlroup status only exposed to one intensity during one state oflertness, were analyzed by means of a repeated measuresNOVA with amplitude as dependent variable, stimulation-site

somatosensory and motor cortex) as between and stimulus (1, 2)s within subjects factor.

ADs were subject to analysis if they fulfilled the followingriteria: oscillations which occurred immediately after stimulationwithin 5 ms after the first pulse), have a spike amplitude of at leastwice the background EEG and a duration of at least 1 s.

Frequency spectra of ADs as recorded next to the stimulationite in either the somatosensory cortex or motor cortex wereenerated for WAG/Rij and Wistar rats. AD spectra of WAG/Rijats were analyzed by means of a repeated measures ANOVAith amplitude (power in a given frequency band) as dependentariable, stimulation site (somatosensory and motor cortex) asetween subjects factor and frequency band (3–5, 7–9, 15–17,3–25 Hz) as within subjects factor. Secondly, the duration ofAG/Rij ADs were analyzed by a 3-way ANOVA with stimulation

ite, intensity as well as state of alertness as between subjectsactors.

Furthermore, the probability of AD occurrence was calculatedor ADs induced with a stimulation intensity of 60 �A for both

AG/Rij rats and Wistar rats. Probability of ADs in WAG/Rij ratsere analyzed with repeated measures ANOVA with stimulationite as between subjects factor and state of vigilance as withinubjects factor.

Lastly, the correlation between N1 amplitude and AD proba-ility was calculated for all rats and for WAG/Rij and Wistar ratseparately, in order to investigate the relationship between thendicators of local excitability (N1 amplitude) and network activityAD probability).

EEG parts with movement artifacts were not included in thenalysis. In addition only animals with verified electrode positions

n the deep layers (IV–VI) of the somatosensory and motor cortexere included in the statistical analyses.

RESULTS

Electrical evoked potentials in WAG/Rij rats

Stimulation-induced evoked potentials in WAG/Rij ratswere mainly visible in the EEG-channel recorded in thedeep layers (IV–VI) of either somatosensory or motor cor-

various components: N1 (latency 15 ms), P1 (latency 23ms), N2 (latency 27 ms), P2 (latency 31 ms) and N3 (40ms). An exemplary evoked potential as measured in thesomatosensory and motor cortex is shown in Fig. 1 (lowerpanel).

The ANOVA of the N1, the most direct measurement oflocal excitability, revealed a significant main effect for stim-ulation site (F(1,15)�5.284; P�0.05): its amplitude washigher in rats stimulated in the deep somatosensory cortexthan in rats stimulated in the deep motor cortex (see Fig. 1upper panel).

In addition, the N1 also showed a main effect of state(F(1,16)�5.294; P�0.05) and a main effect of intensity(N1: F(1,24)�7.303; P�0.01). Post hoc comparisons dem-onstrated lower amplitudes during wakefulness as com-pared to drowsiness and sleep (P=s�0.05) and increasingamplitudes with increasing stimulation intensities (see Fig.2). The main effect of stimulus was not significant(F(1,15)�0.469; P�0.05): the amplitude of N1 was equallyhigh for the first and second pulse (see Fig. 1 lower panel).Neither the stimulation-site�state nor the stimulation-site�intensity interaction turned out to be significant (seeFig. 2) demonstrating that the difference in N1 amplitudebetween somatosensory and motor cortex stimulation wasequally strong at all states of alertness and all testedintensities.

The slower component N3 also displayed a significantmain effect for stimulation site (F(1,15)�7.538; P�0.05).Again the amplitudes were higher in the somatosensorythan in the motor cortex (see Fig. 1 upper panel). Also themain effect of state (F(1,16)�5.703; P�0.05) and intensity(F(1,23)�12.197; P�0.01) were significant. Post hoc com-parisons demonstrated lower amplitudes during wakeful-ness as compared to drowsiness and sleep (P=s�0.05)and increasing amplitudes with increasing stimulation in-tensities (see Fig. 2). The main effect of stimulus was notsignificant (F(1,15)�0.122; P�0.05): the amplitudes of N3were equally high for the first and second pulse (see Fig. 1lower panel). The N3 showed a significant stimulation-site�state interaction (F�7.271; P�0.05). Post hoc testsdemonstrated that stimulation of the somatosensory cortexalways (during all three states of alertness) resulted inhigher amplitudes compared to stimulation of the motorcortex (all P=s�.05), but that the difference between brainstructures was stronger during drowsiness and sleep ascompared to active wakefulness (see Fig. 2).

The N3 displayed a significant stimulation-site�intensity interaction (F�8.808; P�0.01) as well. Post hoctests revealed that stimulation of the somatosensory cortexresulted into higher amplitudes compared to stimulation ofthe motor cortex in all tested intensities (P=s�.05) exceptfor the intensity 20 �A. Furthermore, the difference be-tween brain structures became larger with increasing stim-ulation intensities (see Fig. 2).

The remaining components P1, P2 and N2 neitherhowed a main effect for stimulation-site nor a main effector state, nor a main effect for stimulus (P=s�0.05). N2

S�1 5/10/11 6:09 Art: 12902

174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232

174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231

4.826; P�0.05); its amplitude increased with increasingstimulation intensities.

Electrical evoked potentials in Wistar rats

EEPs in Wistar rats displayed similar waveforms with sim-ilar latencies as were found in WAG/Rij rats (see captionElectrical evoked potentials in WAG/Rij rats).

In contrast to WAG/Rij rats, the ANOVA of the N1 didnot reveal a significant main effect of stimulation-site(F(1,14)�0.511; P�0.487). Amplitudes of the N1 did notdiffer for Wistar rats stimulated in the somatosensory ascompared to Wistar rats stimulated in the motor cortex(Fig. 1 upper panel). The main effect of stimulus and thestimulation-site�stimulus interaction for component N1turned out to be not significant, too.

A significant main effect of stimulation-site was found(F(1,14)�15.615; P�0.01) for N3, as was the case inWAG/Rij rats. N3 showed a significantly higher amplitudefor the somatosensory as compared to the motor cortex,

Fig. 1. Electrical evoked potentials (EEPs) in WAG/Rij and Wistar ratsof the somatosensory cortex (left panel) or the deep layers of the mot�A and were applied during non-REM sleep. ECoG recordings were m(Mean and s.e.m.) of components N1 (left) and N3 (right) for evoked postates of alertness) of either absence epileptic WAG/Rij rats or non-epof 60 �A while WAG/Rij data were averaged across the whole intensitegend, the reader is referred to the Web version of this article.

but this effect was less pronounced than in the WAG/Rij

rats: a difference of 70% between somatosensory andmotor cortex was found for Wistar rats as compared to adifference of 90% for WAG/Rij rats (comparison controlled/corrected for intensity of stimulation) (Fig. 1 upper panel).No other main or interaction effect was found for N3 or anyother EEP component in Wistar rats.

ADs in WAG/Rij rats

ADs were noticed in the EEG of WAG/Rij rats after thepresentation of the double pulse stimuli. In contrast to theEEPs, ADs were always visible in both, the recordingchannel close to the stimulation site and the more remoterecording site (i.e. the motor cortex for rats stimulated inthe somatosensory cortex and the somatosensory cortexfor rats stimulated in the motor cortex). Between 33% and87% of WAG/Rij rats displayed a mean of 14 (range 1–68)ADs, depending on the stimulation site and vigilance state.More details can be found in Table 1.

The morphological appearance of ADs in WAG/Rij rats

anel: Exemplary EEPs induced by stimulation in either the deep layers(right panel) of a WAG/Rij rat. Induction stimuli had an intensity of 40

electrodes near (1 mm) the stimulation site. Upper panel: Amplitudenduced in the somatosensory cortex or motor cortex (averaged acrossistar rats. Note that Wistar rats were only stimulated with an intensity0 �A) range. For interpretation of the references to color in this figure

. Lower por cortexade fromtentials iileptic Wy (20–10

was spike–wave-like, rather similar to spontaneous SWDs.

S�1 5/10/11 6:09 Art: 12902

233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291

233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

Frequency spectra generated via Fast Fourier Transforma-tion showed for both ADs, induced by stimulation in thesomatosensory and motor cortex, a dominant frequency of7–9 Hz as is the case for spontaneous SWDs. In additionto the major peak at 8 Hz, smaller peaks were sometimespresent at the frequency of the first (15–17 Hz) and secondharmonic (23–25 Hz).

The ANOVA on the amplitude of AD spectra revealedsignificant main effect of stimulation site (F(1,11)�8.826;

P�0.05): spectra of ADs induced by stimulation of thesomatosensory cortex had higher amplitudes as comparedto spectra of ADs induced by stimulation of the motorcortex (Fig. 3 left panel). In addition, a main effect of

Fig. 2. Amplitude (Mean and s.e.m.) of components N1 and N3 for somdifferent states of alertness (left) and for different stimulation intensitiereader is referred to the Web version of this article.

Table 1. Number of rats showing stimulation-induced afterdischarges

Group Wakefulness

WAG/Rij stimulated in somatosensory cortex six out of eight�WAG/Rij stimulated in motor cortex five out of nine�

Wistars stimulated in somatosensory cortex three out of eigh

frequency band (F(2,24)�37.01; P�0.01) was found: the7–9 Hz band showed higher amplitudes compared to allother frequency bands. Lastly, an interaction betweenstimulation site and frequency band (F(1,24)�14.66;P�0.01) was found. Post hoc tests showed higher ampli-tude values for the 7–9 Hz band for ADs induced bystimulation in the somatosensory cortex as compared toADs induced by stimulation in the motor cortex (P�.01),but no significant area differences for the other frequencybands (Fig. 3).

In addition, differences in the duration of the ADs werenoticed. The ANOVA revealed a significant main effect ofstimulation site (F(1,560)�17.47; P�.001): ADs lasted lon-

ory as compared to motor cortex induced EEPs of WAG/Rij rats duringFor interpretation of the references to color in this figure legend, the

ree different states of alertness

Drowsiness Non-REM sleep

seven out of eight��87% six out of eight��76%seven out of nine��77% three out of nine��33%— —

atosenss (right).

during th

�75%�55%t��37.5%

— —

S�1 5/10/11 6:09 Art: 12902

ion of the

292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350

292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349

ger when the induction stimulus was given in the somato-sensory compared to stimulation in the motor cortex.Duration of the ADs also depended on the intensity ofstimulation (F(2,560)�50.43, P�.001), intense stimuliproduced longer ADs, and were also dependent on thestate of the animals (F(2,560)�1239.6, P�.0001). Posthoc tests showed that the duration was longer during thedrowsy state compared to wakefulness and sleep. Thestimulation-site�state interaction (Fig. 4, left panel) wasnot significant indicating that the difference in durationbetween somatosensory and motor cortex induced ADswere present irrespective of the state of alertness.

Differences were also seen for the probability of ADinduction: the chance to induce an AD was dependent onthe vigilance state (F(1,18)�10,66; P�0.01). The proba-bility of an AD was higher during drowsiness compared towakefulness and sleep (Fig. 4, right panel). No significanteffect was seen for stimulation site and stimulation

Fig. 3. Afterdischarges (ADs) induced by stimulation. Left panel: ExWAG/Rij rat; exemplary frequency spectrum of a WAG/Rij AD ansomatosensory and motor cortex of WAG/Rij rats. Right panel: Exemprat and exemplary frequency spectrum of a Wistar AD. For interpretatWeb version of this article.

site�vigilance state.

ADs in Wistar rats

Only sporadic and short lasting ADs were seen in four ofthe 16 Wistar rats (Table 1). A representative AD, and afrequency spectrum of such an AD as seen in a Wistar rat,is displayed in Fig. 3, right panel. It illustrates that ADs ofWistar rats are more irregular compared to those of WAG/Rij rats. Likewise, the spectrum of a Wistar AD has no highamplitude peak at 8 Hz as is the case for WAG/Rij rats butpresent double peaks at 7 and 9 Hz, with a medium to lowamplitude. No further statistical analyses were applied tothem considering the small number of Wistar rats display-ing ADs.

Relation between local excitability and networkexcitability

A correlation analysis between N1 amplitude and AD prob-ability for a unified group of WAG/Rij and Wistar rats

AD induced by a stimulation in the deep somatosensory cortex of aamplitude values of the frequency spectra of ADs elicited in theduced by a stimulation in the deep somatosensory cortex of a Wistarreferences to color in this figure legend, the reader is referred to the

emplaryd meanlary AD in

revealed a non significant result (P�0.577, r�0.098). This

S�1 5/10/11 6:09 Art: 12902

Tbssdm6

telba1icmaaga

saymcctvs

351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409

351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408

also held when WAG/Rij rats and Wistar rats were ana-lyzed as separate groups (P�0.455, r�0.19 and P�0.058,r�0.4).

DISCUSSION

The study aimed to investigate whether there is highercortical excitability in the deep somatosensory in compar-ison to the deep motor cortex in WAG/Rij rats, and not innonepileptic control rats, as would be in line with the cor-tical focus theory of absence epilepsy (Meeren et al., 2002;van Luijtelaar and Sitnikova, 2006). Electrical stimulationwas used to assess local excitability, the amplitude of theearly component(s) of the EEP was used as such. Theslower components of the EEPs might reflect cortical net-work or subcortical network activity, while properties ofADs are assumed to predominantly reflect rhythmic corti-cal-subcortical network oscillations (see “Similarities be-tween ADs and SWDs” for argumentation).

Excitability as assessed by electrical evokedpotentials

The major outcome of the present study was the largeramplitude of the N1 in the electrical stimulation elicitedevoked response in the deep layers of the somatosensoryvs. the same response in the motor cortex of WAG/Rij rats.Since the latencies of the N1 but also the other compo-nents of the EEPs, as elicited and measured in both deeplayers, were rather similar they can be directly comparedand interpreted. Given this, the outcome clearly demon-strates a higher excitability of the epileptic zone in thesomatosensory cortex. The non-significant correlation be-tween N1 amplitude and AD probability further demon-strates and ensures that the higher excitability in WAG/Rijsomatosensory cortex is a local property.

Since no difference in N1 amplitude was found in non-pileptic control rats stimulated with 60 �A, it is concluded

Fig. 4. Duration (mean and s.e.m.) (left) and mean probability (right)motor cortex during three behavioral states (wake, drowsiness, non-Rto 87% of WAG/Rij rats that display a mean of 14 (range 1 to 68) ADis referred to the Web version of this article.

that that this heightened excitability represents a charac- e

teristic of an epileptic focus rather than a normal differencein excitability between somatosensory and motor corticesof healthy individuals. Whether Wistar rats might show adifference in N1 EEP-amplitude between somatosensoryand motor cortex when stimulated with a higher intensityof, for example, 100 �A is not clear from this experiment.

his is however unlikely since differences in N1 amplitudeetween somatosensory and motor cortex were equallytrong for all tested intensities in WAG/Rij rats. In additionuch a result would not change our conclusion since aifference in N1 amplitude between somatosensory andotor cortex was found also for WAG/Rij rats stimulated at0 �A but not for Wistar rats.

Measurements of EP amplitude were done duringhree states of alertness in order to investigate whether thepileptic excitability may change during the natural vigi-

ance states and to control for by vigilance induced possi-le confounding factor on EEP amplitude (van Luijtelaar etl., 1998; Meeren et al., 1998; Loewy et al., 2000; Coenen,995). The degree of local hyperexcitability, as expressed

n the amplitude of N1, was independent of state, as can beoncluded from the fact that the difference between so-atosensory and motor cortex was equally strong duringll states of vigilance. Therefore it seems that the highmplitude of the N1 reflects increased cortical excitabilityiven the state independency of the degree of hyperexcit-bility of the somatosensory cortex.

The excitability of the sensory-motor cortex was earliertudied in WAG/Rij rats (Tolmacheva et al., 2004; Maresnd Tolmacheva, 2007). These authors compared pres-mptomatic and symptomatic WAG/Rij rats with ageatched nonepileptic control rats but from their studies no

ompelling evidence was obtained for an increased corti-al excitability considering that both ACI (nonepileptic con-rols) and WAG/Rij rats showed reduced thresholds forarious types of electrical stimulation induced ADs andeizures in comparison to Wistar controls. Various differ-

scharges (ADs) induced by stimulation in the somatosensory and thep). Depending on the vigilance state, data are based on ADs of 33%erpretation of the references to color in this figure legend, the reader

of afterdiEM slees. For int

nces including the electrode location and stimulation pa-

S�1 5/10/11 6:09 Art: 12902

410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468

410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467

rameters explain the difference between the studies. Otherresults in vitro and in vivo are in good agreement with oururrent findings. Those studies hint to the possible under-ying causes of the increased hyperexcitability found forhe somatosensory cortex of WAG/Rij rats including aeduction in hyperpolarization-activated cation current (Ih)

(Strauss et al., 2004; Kole et al., 2007), an upregulation ofsodium channels (Klein et al., 2004; Blumenfeld et al.,2008), increased NMDA receptor-mediated activity (Luh-mann et al., 1995; D’Antuono et al., 2006; Pumain et al.,1992), a decrease in the efficacy of GABAA or GABABergicinhibition (Merlo et al., 2007; Inaba et al., 2009; van Lui-jtelaar and Sitnikova, 2006) and an increased expressionof mGlu2/3 receptors (Ngomba et al., 2005). Therefore, itseems now well established from both in vivo and in vitrostudies that there is indeed an increased excitability ordecreased inhibition in the somatosensory cortex in ge-netic epileptic rats; however, not or less in other parts ofthe cortex.

A difference between somatosensory and motor cortexof WAG/Rij rats was also found for the amplitude of theslower component N3. While the N1 represents local prop-erties (early components are primarily determined by stim-ulus characteristics (Coenen, 1995)), the slower N3 can beseen to reflect network properties, since later EP compo-nents can also be influenced by higher order, top downprocesses via cortico-thalamo-cortical network interactions(Coenen, 1995). The difference in N3 amplitude betweensomatosensory and motor cortex induced EEPs was pres-ent for both rat strains but this difference between somato-sensory and motor cortex was much more pronounced inWAG/Rij rats than in the Wistar rats. This might either bethe result of the increased excitability in WAG/Rij somato-sensory cortex or might point toward additional differencesin cortical or cortico-thalamo-cortical network propertiesand activity involved in shaping the morphology of EEPsbetween the two strains.

Similarities between ADs and SWDs

ADs in the WAG/Rij rat, induced by both, somatosensoryand motor cortex stimulation, showed remarkable similar-ities to spontaneous SWDs: ADs had a spike wave likemorphology with equal amplitudes throughout the wholeoscillation. Furthermore, their frequency spectrum wasrather similar with a major peak around 8 Hz, representingthe dominant frequency of ADs and SWDs, and smallerpeaks at the first and second harmonic (see Sitnikova etal., 2009 for the frequency spectrum of spontaneousSWDs). Moreover, in contrast to the evoked potentials, theADs were clearly present in the remote recording channel.It seems that these 8 Hz ADs quickly generalize as is thecase for SWDs.

Another striking similarity between both epileptic oscil-lations is their strong preference to occur during thedrowsy state. AD duration and probability of occurrenceduring drowsiness was more than three times higher com-pared to the wake and non-REM sleep state. This wasindependent of the location where ADs were elicited, con-

location. A similar strong preference for the drowsinessstate is also well known for the occurrence of the sponta-neous SWD (Drinkenburg et al., 1991). Neither the exis-tence of this type of 8 Hz ADs, nor the large effects ofvigilance on this type of ADs were described earlier.

Given these strong similarities between spontaneousSWD and stimulation-induced ADs it can be assumed thatthe ADs, recorded in this experiment, represent cortico-thalamo-cortical network activity. This might also be sup-ported by the fact that SWD-like oscillations can be foundafter low frequent thalamic stimulation (Mirski et al., 1997)or by single pulse stimulation of the thalamus, which in-duces rhythmic ADs in adult rats (Mares et al., 1982). Inaddition, unpublished data from our group on the effects ofstimulation of the somatosensory cortex combined withmultiple cortical and thalamic local field potentials revealedthat this type of ADs were simultaneously present in cortexand thalamus.

Network activity as assessed by afterdischarges

The ADs are supposed to reflect oscillations in a cortico-thalamo-cortical network (see paragraph Similarities be-tween ADs and SWDs; Mirski et al., 1997; Mares et al.,1982). Some of the effects on ADs were specific for theregion of stimulation, other effects were region indepen-dent. Therefore both local properties of the brain-structureand global properties of the cortico-thalamic network (loop)can be inferred from them.

Location independent effects: All ADs have a strongpreference to occur during the drowsy state and spectra ofall AD showed a dominant frequency of 8 Hz. Since this isalso the case for SWDs (see upper paragraph), it is con-cluded that 8 Hz is the optimal frequency for both types ofepileptic oscillations in the cortico-thalamo-cortical systemgiven the drowsy state. This preferred frequency is notlikely to be driven by the stimulation frequency, consideringthat only two stimuli were presented with an interstimulusinterval of 400 ms (2.5 Hz). Whether stimulation with astimulus train of 8 Hz might further increase the probabilityof ADs during drowsiness needs to be investigated. Thechanges in firing pattern of the pyramidal neurons in thecortico-thalamo-cortical network from a tonic to a burstingfiring mode, which are due to changes in the membranepotential of these cells (more hyperpolarization), mightunderlie the preferred occurrence of ADs during drowsi-ness.

The probability to induce ADs was also independent ofthe stimulation site. This was unexpected since stimulationwithin the hyperexcitable epileptic zone was expected toresult in more ADs. Perhaps stimulation of the motor cortexquickly involves the epileptic zone since high coherencevalues between motor and somatosensory cortex duringSWD suggest strong intrahemispheric coupling (Sitnikovaand van Luijtelaar, 2006). Furthermore, a role in corticalsubcortical interaction cannot be excluded. The lack ofcorrelation in WAG/Rij rats between the amplitude of theN1, indicating local excitability, and AD probability, indicat-

S�1 5/10/11 6:09 Art: 12902

ccRtrulWbDbtaHeKfesmnpcohk2C

469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527

469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526

work, suggests that additional changes exist at a thalamiclevel that favor the occurrence of ADs in WAG/Rij rats.

Also the sparse presence of ADs in Wistar rats with theurrent chosen low intensities testifies for additionalhanges in the cortico-thalamo-cortical network of WAG/ij rats. Whether these network changes responsible for

he absence or presence of ADs are epilepsy related orelated to another non-epileptic strain difference is yetnclear. In an earlier study, Tolmacheva et al. (2004) found

arge differences in the duration of slow 3 Hz ADs betweenAG/Rij and non-epileptic control rats but no difference

etween symptomatic and presymptomatic WAG/Rij rats.uring this period the epileptic phenotype, as expressedy the age dependent increase of SWDs, increased, whilehe duration of the 3 Hz ADs remained constant (Tolm-cheva et al., 2004). This suggests that the duration of 3z ADs are dependent on the strain but not related to thepileptic phenotype. It might well be that 3 (Mares andubova, 2006; Tolmacheva et al., 2004) and 8 Hz ADs, as

ound in the present study, are difficult to compare consid-ring quite some methodological differences between thetudies done in the Mares laboratory and in ours. It re-ains thus possible that the large probability of theseewly discovered 8 Hz ADs in WAG/Rij rats can be ex-lained by epilepsy related changes in the cortico-thalamo-ortical network, including the RTN. Changes in propertiesf GABA and glutamate and their modulators in the RTNave been described in the WAG/Rij rat (van de Boven-amp-Janssen et al., 2006; Ngomba et al., 2005; Liu et al.,007; van Rijn et al., 2010), while an upregulation in T-typea2� channels has been described in GAERS (Tsakiridou

et al., 1995). The lateral thalamus, including the RTN, hasan indispensable contribution in propagation and mainte-nance of SWD (Liu et al., 1992, 2007; Sohal et al., 2003;Meeren et al., 2009), and most likely also of 8 Hz ADs.

Location dependent effects: the 8 Hz ADs elicited atthe somatosensory cortex lasted longer than those in-duced in the motor cortex. This difference might be seenas support for the differences in excitability between thetwo regions. Alternatively, the difference in AD durationmight reflect properties of different sub-loops of the cortico-thalamo-cortical system.

Also spectral differences between somatosensory andmotor cortex elicited ADs were found as indicated by asignificant main effect of stimulation site for the amplitudesof AD spectra. The remarkably larger peak in the 7–9 Hzband of ADs elicited in the somatosensory cortex pointstoward the involvement of a large number of cortical neu-rons firing synchronously in a stable rhythm. The stabilityof the 8 Hz rhythm also helps to explain the longer durationof somatosensory cortex induced ADs as compared tomotor cortex induced ADs. Under the assumption thatcortex and thalamus fire synchronously during SWDs (orshow increased coupling, Sitnikova et al., 2008), a seizureinitiator (cortex) imposing a strong and stable rhythm to therest of the system (thalamus) is less likely to be disruptedby phase desynchronization, and an abortion of the epi-leptic oscillation. Alternatively, the large 8 Hz spectral peak

Hz rhythm within the whole somatosensory loop as com-pared to the motor-loop of the cortico-thalamic system, thelatter serving a less favorable condition for long ADs.

In sum, there is a focally heightened excitability in thedeep layers of the perioral region of somatosensory cortexin WAG/Rij rats, the region from which the SWDs areemerging. The increased local excitability is expressed inthe larger amplitude of the N1. These results are in agree-ment with the cortical focus theory for absence epilepsy.The differential properties of the ADs suggest differencesin cortico-thalamo-cortical circuits between epileptic andnon-epileptic rats. The longer duration and higher proba-bility for the 8 Hz ADs in the drowsy state and the wellknown high prevalence of SWDs during drowsiness dem-onstrates that drowsiness is rather favorable for the occur-rence of epileptic cortico-thalamo-cortical oscillations andsuggests that 8 Hz ADs and SWDs share common networkproperties. Their constituent cells might have the sameneurophysiological properties, such as burst firing andmembrane potential for their occurrence.

Acknowledgments—The authors thank Gerard van Oijen and JosWittebrood for technical assistance; Hans Krijnen and SaskiaHermeling for animal care and Anna Pitas for assistance in dataacquisition. Parts of the project were financed by BrainGain; Dr.Shouwen Zhang was a recipient of a Huijgens Grant.

REFERENCES

Bancaud J (1969) Physiopathogenesis of generalized epilepsies oforganic nature. In: The physiopathogenesis of the epilepsies(Gestaut H, Jasper HH, Bancaud J, Waltregny A, eds), pp 158–185. Springfield: Charles C Thomas Publisher.

Blumenfeld H, Klein JP, Schridde U, Vestal M, Rice T, Khera DS,Bashyal C, Giblin K, Paul-Laughinghouse C, Wang F, Phadke A,Mission J, Agarwal RK, Englot DJ, Motelow J, Nersesyan H, Wax-man SG, Levin AR (2008) Early treatment suppresses the devel-opment of spike-wave epilepsy in a rat model. Epilepsia 49(3):400–409.

Buzsáki G (1991) The thalamic clock: emergent network properties.Neuroscience 41:351–364.

Coenen AM (1995) Neuronal activities underlying the electroenceph-alogram and evoked potentials of sleeping and waking: implica-tions for information processing. Neurosci Biobehav Rev 19:447–463.

D’Antuono M, Inaba Y, Biagini G, D’Arcangelo G, Tancredi V, Avoli M(2006) Synaptic hyperexcitability of deep layer neocortical cells ina genetic model of absence seizures. Genes Brain Behav5:73–84.

Drinkenburg WH, Coenen AM, Vossen JM, van Luijtelaar EL (1991)Spike-wave discharges and sleep-wake states in rats with absenceepilepsy. Epilepsy Res 9:218–224.

Gloor P (1968) Generalized cortico-reticular epilepsies: some consid-erations on the pathophysiology of generalized bilaterally synchro-nous spike and wave discharges. Epilepsia 9:249–263.

Inaba Y, D’Antuono M, Bertazzoni G, Biagini G, Avoli M (2009) Dimin-ished presynaptic GABA(B) receptor function in the neocortex of agenetic model of absence epilepsy. Neurosignals 17(2):121–131.

Klein JP, Khera DS, Nersesyan H, Kimchi EY, Waxman SG, Blumen-feld H (2004) Dysregulation of sodium channel expression in cor-tical neurons in a rodent model of absence epilepsy. Brain Res1000:102–109.

Kole MH, Bräuer AU, Stuart GJ (2007) Inherited cortical HCN1 chan-nel loss amplifies dendritic calcium electrogenesis and burst firing

S�1 5/10/11 6:09 Art: 12902

528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586

528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579

Kostopoulos GK (2000) Spike-and-wave discharges of absence sei-zures as a transformation of sleep spindles: the continuing devel-opment of a hypothesis. Clin Neurophysiol 111 (Suppl 2):S27–S38.

Liu XB, Coble J, van Luijtelaar G, Jones EG (2007) Reticular nucleus-specific changes in alpha3 subunit protein at GABA synapses ingenetically epilepsy-prone rats. Proc Natl Acad Sci U S A 104:12512–12517.

iu Z, Vergnes M, Depaulis A, Marescaux C (1992) Involvement ofintrathalamic GABAB neurotransmission in the control of absenceseizures in the rat. Neuroscience 48(1):87–93.

oewy DH, Campbell KB, de Lugt DR, Elton M, Kok A (2000) Themismatch negativity during natural sleep: intensity deviants. ClinNeurophysiol 111:63–72.

üders H, Lesser RP, Dinner DS, Morris HH (1984) Generalizedepilepsies: a review. Cleve Clin Q 51:205–226.

uhmann HJ, Mittmann T, van Luijtelaar G, Heinemann U (1995)Impairment of intracortical GABAergic inhibition in a rat model ofabsence epilepsy. Epilepsy Res 22:43–51.

ares P, Maresova D, Trojan S, Fischer J (1982) Ontogenetic devel-opment of rhythmic thalamo-cortical phenomena in the rat. BrainRes 8:765–769.

ares P, Kubova H (2006) Electrical stimulation induced models ofseizures. In: Models of seizures and epilepsy (Pitkanen A,Schwartzkroin PA, Moshe SL, eds), pp 153–160. San Diego: Else-vir Academic Press.

ares P, Tolmacheva E (2007) Comparison of cortical epileptic after-discharges in immature genetic absence epilepsy WAG/Rij ratswith those in two other strains (ACI and Wistar). Epilepsia 48(Suppl 5):164–167. Erratum in: Epilepsia (2007) 48(12):2381.

eeren HK, van Luijtelaar EL, Coenen AM (1998) Cortical and tha-lamic visual evoked potentials during sleep-wake states and spike-wave discharges in the rat. Electroencephalogr Clin Neurophysiol108:306–319.

eeren HK, Pijn JP, van Luijtelaar EL, Coenen AM, Lopes da Silva FH(2002) Cortical focus drives widespread corticothalamic networksduring spontaneous absence seizures in rats. J Neurosci 22:1480–1495.

eeren H, van Luijtelaar G, Lopes da Silva F, Coenen A (2005)Evolving concepts on the pathophysiology of absence seizures:the cortical focus theory. Arch Neurol 62:371–376.

eeren HK, Veening JG, Möderscheim TA, Coenen AM, van LuijtelaarG (2009) Thalamic lesions in a genetic rat model of absenceepilepsy: dissociation between spike-wave discharges and sleepspindles. Exp Neurol 217:25–37.

erlo D, Mollinari C, Inaba Y, Cardinale A, Rinaldi AM, D’Antuono M,D’Arcangelo G, Tancredi V, Ragsdale D, Avoli M (2007) ReducedGABAB receptor subunit expression and paired-pulse depressionin a genetic model of absence seizures. Neurobiol Dis 25:631–641.

irski MA, Rossell LA, Terry JB, Fisher RS (1997) Anticonvulsanteffect of anterior thalamic high frequency electrical stimulation inthe rat. Epilepsy Res 28(2):89–100.

gomba RT, Biagioni F, Casciato S, Willems-van Bree E, Battaglia G,Bruno V, Nicoletti F, van Luijtelaar EL (2005) The preferentialmGlu2/3 receptor antagonist, LY341495, reduces the frequency ofspike-wave discharges in the WAG/Rij rat model of absence epi-lepsy. Neuropharmacology 49:89–103.

iedermeyer E (1972) The generalized epilepsies: a clinical electroen-cephalographical study. Springfield: Charles C Thomas Publisher.

axinos G, Watson C (1998) The rat brain in stereotactic coordinates.

enfield WG, Jsper HH (1954) Epilepsy and the functional anatomy ofthe human brain. Boston, MA: Little Brown & Co.

olack PO, Guillemain I, Hu E, Deransart C, Depaulis A, Charpier S(2007) Deep layer somatosensory cortical neurons initiate spike-and-wave discharges in a genetic model of absence seizures.J Neurosci 27(24):6590–6599.

olack PO, Mahon S, Chavez M, Charpier S (2009) Inactivation of thesomatosensory cortex prevents paroxysmal oscillations in corticaland related thalamic neurons in a genetic model of absence epi-lepsy. Cerebral Cortex 19(9):2078–2091.

umain R, Louvel J, Gastard M, Kurcewicz I, Vergnes M (1992)Responses to N-methyl-D-asparate are enhanced in rats with petitmal-like seizures. J Neural Transm Suppl 35:97–108.

itnikova E, van Luijtelaar G (2006) Cortical and thalamic coherenceduring spike-wave seizures in WAG/Rij rats. Epilepsy Res 71(2–3):159–180.

itnikova E, Dikanev T, Smirnov D, Bezruchko B, van Luijtelaar G(2008) Granger causality: cortico-thalamic interdependencies dur-ing absence seizures in WAG/Rij rats. J Neurosci Methods170(2):245–254.

itnikova E, Hramov AE, Koronovsky AA, van Luijtelaar G (2009)Sleep spindles and spike-wave discharges in EEG: their genericfeatures, similarities and distinctions disclosed with Fourier trans-form and continuous wavelet analysis. J Neurosci Methods 180:304–316.

ohal VS, Keist R, Rudolph U, Huguenard JR (2003) DynamicGABA(A) receptor subtype-specific modulation of the synchronyand duration of thalamic oscillations. J Neurosci 23:3649–3657.

teriade M (2003) Neuronal substrates of sleep and epilepsy. Cam-bridge: Cambridge University Press.

trauss U, Kole MH, Bräuer AU, Pahnke J, Bajorat R, Rolfs A, NitschR, Deisz RA (2004) An impaired neocortical Ih is associated withenhanced excitability and absence epilepsy. Eur J Neurosci 19:3048–3058.

olmacheva EA, van Luijtelaar G, Chepurnov SA, Kaminskij Y, MaresP (2004) Cortical and limbic excitability in rats with absence epi-lepsy. Epilepsy Res 62:189–198.

sakiridou E, Bertollini L, de Curtis M, Avanzini G, Pape HC (1995)Selective increase in T-type calcium conductance of reticular tha-lamic neurons in a rat model of absence epilepsy. J Neurosci15(4):3110–3117.

an de Bovenkamp-Janssen MC, van der Kloet JC, van Luijtelaar G,Roubos EW (2006) NMDA-NR1 and AMPA-GluR4 receptor sub-unit immunoreactivities in the absence epileptic WAG/Rij rat. Epi-lepsy Res 69(2):119–128.

an Luijtelaar EL, Miller CA, Coenen AM, Drinkenburg WH, EllenbroekBA (1998) Differential effects of non-REM and REM sleep onsensory gating in rats. Acta Neurobiol Exp 58:263–270.

an Luijtelaar G, Sitnikova E (2006) Global and focal aspects ofabsence epilepsy: the contribution of genetic models. NeurosciBiobehav Rev 30:983–1003.

an Luijtelaar G, Sitnikova E, Lüttjohann A (2011) On the origin andsuddenness of absences in genetic absence models. Clin EEGNeurosci 42:1–5.

an Rijn CM, Gaetani S, Santolini I, Badura A, Gabova A, Fu J,Watanabe M, Cuomo V, van Luijtelaar G, Nicoletti F, Ngomba RT(2010) WAG/Rij rats show a reduced expression of CB1 receptorsin thalamic nuclei and respond to the CB1 receptor agonist,R(�)WIN55,212-2, with a reduced incidence of spike wave dis-

580581582583584585

(Accepted 14 April 2011)

S�1 5/10/11 6:10 Art: 12902

Electrical stimulation of the epileptic focus in absence epileptic WAG/Rij rats: assessment of local...

Documents

Prediction of epileptic seizures

Acquired Etiological Factors in 1,785 Epileptic Subjects: Clinical-Anamnestic Research

Pericardial Injection of Kainic Acid Induces a Chronic Epileptic

Adaptive epileptic seizure prediction system

Mitochondrial dysfunction in neurological disorders with epileptic phenotypes

Epileptic Seizures Detection Using Deep Learning Techniques

Granger causality: Corticothalamic interdependencies during absence seizures in WAG/Rij rats

Optical imaging of acute epileptic networks in mice

Localization of Epileptic Auras Induced on Stimulation by Subdural Electrodes

Effects of some neurosteroids injected into some brain areas of WAG/Rij rats, an animal model of generalized absence epilepsy

Multimodal epileptic seizure detection: towards a wearable

Hippocampal neurons and glia in epileptic EL mice

Optimization of WAG in Smart Wells: An Experimental Design

Orexins cause epileptic activity

Magnetic Stimulation Can Modulate Seizures in Epileptic Patients

Epileptic Seizure Prediction based on Ratio and Differential Linear Univariate Features

Non-Epileptic Psychogenic Seizures

A) 3-Phase Locomotives ( WAP-7, WAG-9) 1. 3phase

Parvalbumin and calbindin-D28k immunocytochemistry in human neocortical epileptic foci

Extracellular metabolites in the cortex and hippocampus of epileptic patients