The electrophysiological "delayed effect" of focal interictal epileptiform discharges: a low resolution electromagnetic tomography (LORETA) study

Epilepsy Resea ch 2009;85:270-278. The elec ophysiological "delayed e ec " o ocal in e ic al epilep i o m discha ges. A low esolu ion elec omagne ic omog aphy (LORETA) s udy *1 Clemens Béla M.D., Ph.D.; 1 Pi os Pálma M.D.; 1 Bessenyei Mónika M.D.; 2 Va ga Edi , M.D.; 3 Puskás Szil ia, M.D., 3 Feke e Is án, M.D., Ph.D. 1 Kenézy Hospi al L d., Depa men o Neu ology, Deb ecen, Hunga y 2 Kenézy Hospi al L d., Depa men o Pedia ics, Deb ecen, Hunga y 3 Uni e si y o Deb ecen, Medical and Heal h Science Cen e , Depa men o Neu ology, Deb ecen, Hunga y * Co esponding au ho : D . Clemens Béla, Kenézy Kó ház K ., Neu ológia, Ba ók Béla ú 3., 4031, Deb ecen, Hunga y. TEL: ++36 52 511 777 Fax: ++36 51 511 729 E-mail: [email protected] Abs ac Colla ing he indings ega ding he ole o ocal in e ic al epilep i o m discha ges (IEDs) on CNS unc ions aises he possibili y ha IEDs migh ha e nega i e impac ha ou las s he du a ion o he spike-and wa e complexes. The aim o his s udy was he elec ophysiological demons a ion o he "delayed e ec " o he IEDs. 19-channel, linked-ea s e e enced, digi al waking EEG eco ds o ele en child en (aged 6-14 yea s, eigh wi h idiopa hic, h ee wi h c yp ogenic ocal epilepsy, showing a single spike ocus) we e e ospec i ely selec ed om ou da abase. A minimum o wen y (p e e ably, hi y), 2-second epochs con aining a single ocal spike-and-wa e complex we e selec ed (Spike epochs). The ea e , Pos spike-1 (Ps1), Pos spike-2 (Ps2) and Pos spike-3 (Ps3) epochs we e selec ed, ep esen ing he i s and second seconds (Ps1), he hi d and ou h seconds (Ps2) and he i h and six h seconds (Ps3) a e he Spike epoch, espec i ely. In e spike epochs (Is) we e selec ed a a dis ance a leas en seconds a e he Spike epoch. Indi idual analysis: he equency o in e es (FOI = he indi idual equency o he wa e componen o he IEDs), and he egion o in e es (ROI = he si e o he IEDs) we e iden i ied by eading he aw EEG wa e o m and he ins an powe spec um. Ve y na ow-band LORETA (low esolu ion elec omagne ic omog aphy) analysis a he FOI and ROI was ca ied ou . Age-adjus ed, Z- ans o med LORETA "ac i i y" (= cu en sou ce densi y, ampe s / me e s squa ed) was compa ed in he Spike, Ps1, Ps2, Ps3 and Is epochs. Findings: he g ea es (uppe mos pa hological) Z-sco es and he g ea es spa ial ex ension o he LORETA-abno mali y we e always ound in he Spike epochs, ollowed by he g adual dec ease o ac i i y in e ms o se e i y and spa ial ex ension in he Ps1, Ps2, Ps3 epochs. The lowes (baseline) le el and ex ension o he abno mali y was ound in he Is epochs. G oup analysis: a e age alues o ac i i y ac oss he pa ien s we e compu ed o he empo al dec ease o he abno mali y. Findings: a clea endency o he dec ease o abno mali y was demons a ed. Conclusion: he "delayed e ec " o he IEDs was demons a ed elec ophysiologically and quan i ied. The me hod may be u ilized in he indi idual assessmen o he e ec o IEDs on co ical ac i i y, he deg ee and empo o-spa ial ex ension o he abno mali y. Key wo ds Epilepsy, spike, delayed e ec o he spike, EEG, LORETA 1. In oduc ion The elec oencephalog aphic e m in e ic al epilep i o m discha ge (IED) means spikes and sha p wa es usually ollowed by a slow wa e (Cha ian e al, 1974). The a ibu e "in e ic al" indica es ha , unlike ic al EEG pa e ns, IEDs a e no accompanied by seizu e symp oms. Howe e , IEDs a e no always clinically silen because e en a single, ocal spike-and-slow wa e complex migh in e e e wi h a en ional, mo o , senso y and cogni i e p ocesses (Shewmon and E win 1988a-c). The designa ion "noci e ous co ex" implies ha IEDs hemsel es impai neu ological unc ions; in ac , he spiking co ex gi es ise o a g ea e deg ee o ce eb al dys unc ion han he aboli ion o he same pa o he co ex (Pen ield, 1952). In a ecen summa y o he IED- ela ed dys unc ion he au ho concluded ha " ocal IED should no be assumed as clinically benign" (Fisch, 2003). As a co olla y, "EEG ea men ", ha is, aboli ion o he IEDs wi h an iepilep ic d ugs has been a ma e o inc easing in e es (Gue ini e al, 2002; Binnie, 2003; P essle e al, 2005; Jaseja, 2006). The mechanism and he e olu ion o he IED- ela ed de ici symp oms a e no clea ly unde s ood. Some symp oms like ansien aboli ion o muscle one (Rubboli e al, 1995), inc eased eac ion ime (Shewmon and E win, 1988a), and ansien impai men o cogni i e unc ions (Aa s e al, 1984; Binnie and Ma s on, 1992; Aldenkamp e al, 2005) a e s ic ly ime-locked o he IEDs. Howe e , o he in es iga o s desc ibed luc ua ing neu ological symp oms, dis u bed speech and cogni i e impai men ha we e ela ed o he p esence o equen IEDs bu no s ic ly ime-locked o hem. D ug ea men ha abolished he spikes caused signi ican clinical imp o emen wi hin a sho ime pe iod (Hi sch e al, 1990; Me z-Lu z e al, 1999; Baglie o e al, 2001; Massa e al, 2001; Dalla Be na dina e al, 2002). Thus, a "delayed e ec " o he IEDs migh be deduc i ely pos ula ed as an explana ion o he lack o he p ecise empo al coincidence be ween he IEDs and he ime cou se o he abo e men ioned symp oms. We a e no awa e o any explici hypo hesis add essing his delayed e ec albei a ew possible mechanisms ha e been concei ed (Holmes and Lenck-San ini, 2006). The delayed e ec migh be due o oscilla ions ha su i e he spike-and-wa e complex. Con en ional isual EEG analysis con i ms slow EEG oscilla ions a e a single spike in some pa ien s bu no in all. Howe e , quan i a i e assessmen o he impac o hese oscilla ions on physiological co ical ac i i y is no possible by isual analysis. Ou hypo hesis was ha IED- ela ed abno mal oscilla ions ollow he spike-and-slow wa e complex, and he ime cou se and spa ial dis ibu ion o his phenomenon may be cha ac e ized in a quan i a i e way by sou ce analysis. 2. Pa ien s and me hods 2.1. Pa ien s and EEG eco ding The design o his e ospec i e s udy was app o ed by he Resea ch E hics Commi ee o Kenézy Kó ház L d. Ou compu e ized Epilepsy & EEG da abase was e ospec i ely sea ched o epilep ic child en who ul illed he ollowing c i e ia: egis e ed in he 2002 o 2007 ime pe iod; idiopa hic o c yp ogenic, localiza ion- ela ed epilepsy o childhood as de ined by he in e na ional classi ica ion (ILAE 1989); age o onse be ween 6 and 14 yea s; no ea men a EEG e alua ion; EEG ca ied ou a a day wi hou p io seizu e; a e ie able waking EEG eco d ha was eco ded in ou labo a o y wi h he usual echnique. This means he use o he same digi al EEG equipmen , he 10-20 elec ode sys em, eco ding om 19 ac i e elec odes agains he linked ea s e e ence, impedances < 10 kOhm, il e s se a 0.1 and 33.6 Hz, 128 pe second sampling, 2 12 bi on-line analog-digi al con e sion. EEG aces ha did no i he gene al quali y c i e ia o quan i a i e EEG analysis (Nuwe e al, 1994) we e excluded wi h he ema kable di e ence ha he p esence o IEDs was no an exclusion c i e ion. Only he "eyes closed" po ion o he EEG eco ds showing a single spike ocus and IEDs o "medium equency" we e analyzed because e y a e o oo equen spiking did no pe mi he applica ion o he s udy design p esen ed below and in Fig-1. 2.2. Quan i a i e EEG analysis Epoch selec ion and compu ing he Fas Fou ie T ans o m (FFT) we e done wi h he Neu oGuide so wa e (h p://www.appliedneu oscience.com). As he i s s ep, a leas wen y (p e e ably, hi y) 2- second epochs o EEG ac i i y each con aining a single spike-and slow complex we e selec ed as p esen ed in Fig-1 (Spike epochs). The ea e , 2-second epochs immedia ely ollowing he Spike epochs we e selec ed (Pos spike-1 epochs, Ps1), ollowed by he selec ion o Pos spike-2 (Ps2) and Pos spike-3 (Ps3) epochs each o 2 seconds. Finally, 2-second epochs ha we e loca ed a leas 10 seconds om he las Spike epoch (In e spike epochs, Is) we e selec ed. Files selec i ely con aining he Spike, Ps1, Ps2, Ps3, and Is epochs o each pa ien we e p ocessed sepa a ely. FFT alues o he selec ed epochs we e a e aged o e e y ile, and he esul s we e con e ed o ".lo " iles in o de o ans e he da a o he allied LORETA (low esolu ion elec omagne ic omog aphy) so wa e. LORETA is a ecen ly de eloped me hod o localize mul iple dis ibu ed co ical sou ces o bioelec ic ac i i y in he h ee-dimensional space (Pascual-Ma qui e al, 1994). In o he wo ds, LORETA demons a es he synch onously ac i a ed neu onal popula ions unde lying EEG ac i i y by compu ing hei co ical localiza ion om he scalp dis ibu ion o he elec ic ield. This is called sol ing he in e se p oblem o he EEG. The LORETA in e se solu ion is based on exis ing neu oana omical and physiological knowledge and a ma hema ical cons ain called he smoo hness assump ion (Pascual-Ma qui, 2002a). LORETA compu es he in e se solu ion wi hin a h ee-shell sphe ical head model including scalp, skull, and b ain. The b ain compa men o his model was es ic ed o he co ical g ey ma e and hippocampus, acco ding o he Talai ach B ain A las digi ized a Mon eal Neu ological Ins i u e (Talai ach and Tou noux, 1988). The g ey ma e compa men was subdi ided in 2394 oxels, which allows a spa ial esolu ion o 7 millime e s. LORETA compu es a physically exis ing dimension, cu en densi y (ampe s / me e s squa ed) o each oxel. Fo he sake o b e i y, his is called "ac i i y" in his pape . The consis ency o LORETA wi h physiology and localiza ion has been alida ed o a lo o no mal and pa hological condi ions (Pascual-Ma qui e al, 2002b). The accu acy o LORETA in localizing IEDs has been add essed in a ew pape s. In ac anial EEG and unc ional MRI con i med he LORETA-de ined localiza ion o IEDs (Seeck e al, 1998). LORETA-MRI usion disclosed conside able opog aphical o e lap be ween he epilep ogenic lesion and he IEDs (Sgou os e al, 2001). Using 22 scalp elec odes and in ac anial leads LORETA ound he highes cu en densi y in he a ea whe e in ac anial EEG de ec ed IEDs (Lan z e al, 1997). Using 21 scalp elec odes LORETA localized co ec ly IEDs o he mesial empo al s uc u es in 14/19 pa ien s in whom he localiza ion was con i med by simul aneous o amen o ale eco dings and he good esul s o amygdalo-hippocampec omy. The e was no clea LORETA solu ion in he emaining 5 cases bu no LORETA solu ion was clea ly inco ec (Zums eg e al, 2005). Using 23 o 29 scalp elec odes LORETA localized a dis inc EEG abno mali y, small sha p spikes, in he same sub-loba a ea han did in ac anial EEG eco dings (Zums eg e al, 2006). Howe e , he LORETA me hod has i s limi a ions. Comp ehensi e e alua ion o i s localizing accu acy and he scien i ic backg ound o LORETA we e discussed in ecen pape s (Zums eg e al, 2005; G o a e al, 2006; Plumme e al, 2008). Gi en ha he main sou ce o IED- ela ed de ici is he slow wa e componen o he ocal spike-and-slow complex, e y na ow band LORETA analysis (wi h he esolu ion o 1 Hz) was ocused o he equency o in e es (FOI) and he egion o in e es (ROI). FOI and ROI we e i s es ima ed by e alua ing he aw EEG ace and he ins an powe spec um o he a e aged spike epochs. The main nega i e phase o he spikes was i ed o he middle o he spike epoch. (Fig-2). As he nex s ep, FOI (= he equency o he wa e componen o he IEDs) and ROI (= he si e o maximum LORETA abno mali y in he a e aged Spike epochs a he si e o he IEDs) we e p ecisely de e mined om he LORETA solu ion. The ROI was ana omically localized in he h ee-dimensional space by speci ying he posi ion o he oxel on he Z-axis o he Talai ach coo dina e sys em, he numbe o he B odmann-a ea, and he ana omical name o he gy us. Measu emen was ca ied ou a he selec ed FOI and ROI (Fig-3). In his s udy we used age-adjus ed, Z- ans o med LORETA alues o ac i i y he de i a ion o which was desc ibed in he LORETA No ma i e EEG Da abase (Tha che e al, 2005). The use o s a is ical mapping was epo ed o be supe io o he aw LORETA esul s in localiza ion s udies (Zums eg e al, 2005). The deg ee o abno mali y (Z-sco e) wi hin he ROI was compa ed among he Spike, Ps1, Ps2, Ps3 and Is epochs o each pa ien hus demons a ing he se e i y and ime cou se o he abno mali y. Pic u e-mon ages we e c ea ed o each pa ien in o de o assess he spa ial changes o he abno mali y ac oss he Spike, Ps1, Ps2, Ps3 and Is epochs (Figu es 4. and 5). The ana omical ex ension o he abno mali y could be semi- quan i a i ely es ima ed om hese LORETA se ies. Finally, he Z-sco es o he Spike, Ps1, Ps2, Ps3 and Is epochs we e a e aged ac oss he pa ien s in o de o assess he a e age deg ee o empo al dec ease o he abno mali y. 3. Resul s 3.1. Tempo al diminu ion o he IED- ela ed abno mali y The da a o he pa ien s, he EEG localiza ion o he IEDs and he indings a e gi en in Table-1. All he pa ien s displayed maximum abno mali y in he Spike epochs ha showed a ma ked endency o diminish in he Ps1, Ps2, Ps3 epochs and eached he minimum alue in he Is epochs. This endency was ound in he child en wi h idiopa hic pa ial epilepsy wi h olandic spikes (BERS) and in c yp ogenic epilepsy alike. Howe e , he indi idual ime cou se o he dec ease o he abno mali y di e ed om pa ien o pa ien . As o show he a e age endency he da a o he pa ien s we e a e aged. Pai ed - es s showed s a is ically conside able di e ence be ween he Spike and Ps1 epochs (p<0.0001), be ween Ps1 and Ps2 (p=0.004) bu no be ween Ps2 and Ps3 (p=0.276) and Ps3 and Is epochs (p=0.070). The eliabili y o he indings was suppo ed by he high es - e es and spli - hal eliabili y alues. These eliabili y measu es o he 19 de i a ions we e a e aged and he a e age alues we e always abo e 90 pe cen o he Spike, Ps1, Ps2, Ps3 and Is epochs. 3.2. Spa ial ex ension o he abno mali y LORETA always la e alized he maximum o he abno mali y o he hemisphe e o he spike ocus. The e was no s iking disc epancy be ween he LORETA solu ion and he con en ional EEG localiza ion o he IEDs. The LORETA se ies ha we e composed o each pa ien showed ha he spa ial ex ension o he abno mali y was he g ea es in he Spike epochs and showed an o e all endency o g adual spa ial sh inking in he Ps1, Ps2, Ps3 and Is epochs as demons a ed by he wo exempla s (Figu es 4, 5). The Z-sco e helps o es ima e he ex ension and he ime cou se o he abno mali y on he indi idual basis. 4. Discussion 4.1. The impo ance and he assessmen o "synch oniza ion" The IED- ela ed de angemen o co ical unc ions is mainly due o he wa e componen o he discha ge. The wa e ep esen s su ound inhibi ion, he empo o-spa ial dis ibu ion o which de e mines he local and emo e de ici symp oms including cogni i e impai men (P ince and Wilde , 1967; Shewmond and E win, 1988c). E e y EEG eade knows ha he wa es in e up no mal b ain hy hms o a longe ime pe iod (se e al hund ed milliseconds) han he spikes. In addi ion, he wa e is ime-locked o he dec ease o he e y as (100 o 500 Hz) oscilla ions ha a e in ol ed in no mal in o ma ion p ocessing (U es a azu e al, 2006). Ea lie s udies disclosed ha he dis up i e e ec o a single IED is di ec ly p opo ional o he ol age o he wa e bu no he spike (Shewmon and E win, 1988c). This obse a ion and he di ec ela ionship be ween he spa io- empo al amoun o he spike- wa e se ies and he se e i y o he de ici symp oms in he long un (Me z-Lu z e al, 1999; Baglie o e al, 2001; Massa e al, 2001) a e independen ly de i ed indings bu lead o he same conclusion: he g ea e he deg ee o neu onal synch oniza ion o e ime and space he g ea e he co ical dys unc ion. As o a oid con usion in his pape "synch oniza ion" e e s o he synch onous ac i a ion o g ea co ical neu onal masses (Nunez, 1995), he esul o which is e lec ed by he inc ease o he ol age o he EEG signal (P u schelle and Lopes da Sil a, 1999). This ela ionship highligh s why LORETA is an app op ia e me hod o assess IED- ela ed synch oniza ion: " he LORETA in e se solu ion co esponds o he h ee-dimensional dis ibu ion o elec ic neu onal ac i i y ha has maximum simila i y (i.e., maximum synch oniza ion) in e ms o s eng h and o ien a ion be ween neighbou ing neu onal popula ions" (Pascual-Ma qui, 2002a). The use o 2-second epochs was p ede e mined by he so wa e; howe e , i was a lucky choice because his ime ame was su icien o include all componen o he IED including some possible p e-spike e en s (Hawco e al, 2007), and he comple e IED. 4.2. The delayed EEG e ec Ou indings con i med he hypo hesis ha he delayed EEG e ec las s o a leas a ew seconds beyond he du a ion o he IED, and he cha ac e is ics o he dec easing oscilla ion can be quan i a i ely assessed in space and ime. I s neu onal basis is no known bu se e al mechanisms exis ha can gene a e and sus ain hy hmic phenomena in epilep ically unc ioning neu onal ensembles (de Cu is and A anzini, 2001; McCo mick and Con e as, 2001). Nume ically, he deg ee and ime cou se o he delayed EEG e ec e e s o he in es iga ed coho o pa ien s wi h medium spike equency. I is possible ha mo e equen IEDs o long se ies o hem esul in mo e delayed no maliza ion o co ical ac i i y hus esul ing in less ansien CNS dys unc ion. This possibili y is suppo ed by ano he , IED- ela ed phenomenon, he dec ease o axonal i ing o hippocampal neu ons ha ou las s he IED o abou 2 seconds bu e en o longe pe iods a e clus e s o spikes (Zhou e al, 2007). As a as we know, no o he elec ophysiological phenomena ha ou las he en i e spike- and-wa e complex ha e been desc ibed ye . 4.3. Spa ial assessmen o he abno mali y The localiza ion o he IED- ela ed dys unc ion can be isualized by means o quan i a i e EEG me hods, magne o-encephalog aphy and unc ional magne ic esonance imaging ( MRI). The la e me hod g ea ly con ibu ed o he unde s anding he complexi y o he IED- ela ed abno mali y, he in ol emen o subco ical s uc u es ( ha emain hidden o scalp EEG), and di e en ia e be ween a eas o inc eased and dec eased ac i i y. EEG- MRI co- egis a ion seems o be pa icula ly p omising (G o a e al, 2008); howe e , going in o de ails is beyond he scope o his pape . LORETA localized he abno mali y conco dan ly o he aw EEG wa e o m and he EEG-spec um hus inc easing he numbe o he pape s alida ing he LORETA me hod (Pascual-Ma qui e al, 2002b). In his pape we in es iga ed he localiza ion o he sum o he IED- ela ed, spa ially dis ibu ed abno mali y in 2-sec ime ames; his ask undamen ally di e s om he e o s o localize he sou ces o spikes in BERS (o any o he epilepsy synd ome) and mus no be con used wi h hem. No su p isingly he maximum spa ial ex ension o he abno mali y was ound in he Spike epochs and a g adual sh inking o i was seen in he subsequen epochs. The exac neu onal mechanism o his spa ial dec ease is no known bu he ole o dec eased in aco ical synch oniza ion seems o be p obable (Chagnac-Ami ai and Conno s, 1989; de Cu is and A anzini, 2001). Finally, e en he Is epochs show some deg ee o ocal abno mali y, as compa ed o he emaining pa s o he co ical man le. This is in acco d wi h a ecen in ac anial EEG s udy demons a ing ha locally inc eased synch oniza ion in he in e ic al s a e is cha ac e is ic o he epilep ic co ex in ocal epilepsy (Sche on e al, 2007). The sho coming o he s udy is subop imal spa ial sampling due o he limi ed numbe o elec odes, which is usually c i icized in 19-channel LORETA s udies. Ne e heless, we do no hink ha his esul ed in signi ican ly inaccu acy in his s udy. Res ing EEG hy hms (in pa icula , slow hy hms) a e gene a ed by la gely dis ibu ed co ical sou ces ha can be accu a ely in es iga ed by he s anda d 10-20 sys em (Nunez, 1995; Babiloni e al, 2006). Compa a i e s udies disclosed ha LORETA analysis esul s in e y simila localiza ion o he gene a o s wi h 19 and 46 elec odes p o ided ha hey a e e enly dis ibu ed on he scalp (Michel e al, 2004). In ou s udy he good signal o noise a io imp o ed he accu acy o localiza ion. Albei some mislocaliza ion and blu ing he ex ension o he abno mali y may be inhe en o LORETA, his mainly e e s o smalle sou ces (Zums eg e al 2005). In any case, he po en ial localiza ion inaccu acy o he me hod does no in e e e wi h ou main indings. Finally, he LORETA no ma i e da abase ( ha pe mi s he assessmen o he abno mali y in e ms o age- ela ed de ia ions om Z=0 o each oxel) is cons uc ed o 19 s anda d de ia ions and inse ing u he de i a ions is no possible. 0nly he ana omical dis ibu ion a he equency o he wa e was in es iga ed in his s udy. Howe e , he complex, mu ual in e dependence o he physiological oscilla ions a di e se equencies in he b ain (Buzsáki and D aguhn, 2004) implies ha IEDs may cause in e e ence wi h some o he oscilla ions unde lying cogni ion and o he co ical ac i i ies. The p esen indings sugges ha LORETA migh be a use ul me hod o in es iga e he ana omical dis ibu ion o he e ec o dis u bing ansien e en s on no mal ce eb al ac i i ies. 5. Re e ences Aa s, J.H.P., Binnie, C.D., Smi , A.M., Wilkins, A.J., 1984. Selec i e cogni i e impai men du ing ocal and gene alized epilep i o m EEG ac i i y. B ain 107, 293-308. Aldenkamp, A.P., Bei le , J., A ends, J., an de Linden, I., Diepmen, L., 2005. Acu e e ec s o subclinical epilep i o m EEG discha ges on cogni i e ac i a ion. Func ional Neu ology 20, 23-28. Babiloni, C., F isoni, G., S e iade, M., B esciani, L., Bine i, G., Del Pe cio, C., Ge oldi, C., Miniussi, C., Nobili, F., Rod iguez, G., Zappasodi, F., Ca agna, T., Rossini, P.M., 2006. F on al whi e ma e olume and del a sou ces nega i ely co ela e in awake subjec s wi h mild cogni i e impai men and Alzheime 's disease. Clin. Neu ophysiol. 117, 1113-1129. Baglie o, M.G., Ba aglia, F.M., Nobili, L., To o elli, S., De Neg i, E., Cale o, M.G., Veneselli, E., De Neg i, M., 2001. Neu opsychological diso de s ela ed o in e ic al epilep ic discha ges du ing sleep in benign epilepsy o childhood wi h cen o empo al o olandic spikes. De . Med. Child Neu ol. 43, 407- 412. Binnie, C.D., Ma s on, D., 1992. Cogni i e co ela es o in e ic al discha ges. Epilepsia 33 (Suppl. 6), S11-17. Binnie, C.D., 2003. Cogni i e impai men du ing epilep i o m discha ges: is i e e jus i iable o ea he EEG? Lance Neu ol. 2, 725-730. Buzsáki, G., D aguhn, A., 2004. Neu onal oscilla ions in co ical ne wo ks. Science 304, 1926-1929. Chagnac-Ami ai, Y., Conno s, B.W., 1989. Ho izon al sp ead o synch onized ac i i y in neoco ex and i s con ol by GABA-media ed inhibi ion. J. Neu ophysiol. 61, 747-758. Cha ian, G.E., Be gamini, L., Dondey, M., Lennox- Buch al, I., Pe e sen, I., 1974. A glossa y o e ms mos commonly used by clinical elec oencephalog aphe s. Elec oenceph. Clin. Neu ophysiol. 37, 538- 553. Dalla Be na dina, B., Sg o, V., Feje man, N. Epilepsy wi h cen o- empo al spikes and ela ed synd omes. 2002. In: Roge , J., Bu eau, M., D a e , Ch., Gen on, P., Tassina i, C.A., Wol , P. (Eds). Epilep ic synd omes in in ancy, childhood and adolescence. 3nd Edi ion. John Libbey & Co., UK., pp. 181.202. de Cu is, M., A anzini, G., 2001. In e ic al spikes in ocal epilep ogenesis. P og ess Neu obiol. 63, 541-567. Fisch, B.J., 2003. In e ic al epilep i o m ac i i y, diagnos ic and beha io al implica ions: 2002 ACNS P esiden al Add ess. J. Clin. Neu ophysiol. 20, 155-162. G o a, C., Daunizeau, J., Lina, J-M., Béna , C.G., Benali, H., Go man, J., 2006. E alua ion o EEG localiza ion me hods using ealis ic simula ions o in e ic al spikes. Neu oImage, 29, 734-753. G o a, C., Daunizeau, J., Kobayashi, E., Bagshaw, A.P., Lina, J-M., Dubeau, F., Go man, J., 2008. Conco dance be ween dis ibu ed EEG sou ce localiza ion and simul aneous EEG- MRI s udies o epilep ic spikes. Neu oImage 39, 755-774. Gue ini, R., A zimanoglu, A., B ouwe , O., 2002. Ra ionale o ea ing epilepsy in child en. Epilep ic Diso de s 4 (Suppl. 2), S9-S21. Hawco, C.S., Bagshaw, A.P., Lu, Y., Dubeau, F., Go man, J., 2007. BOLD changes occu p io o epilep ic spikes seen on scalp EEG. Neu oImage 35, 1450-1458. Hi sch, E., Ma escaux, C., Maque , P., Me z-Lu z, M.N., Kiesmann M., Salmon, E., F anck, G., Ku z, D., 1990. Landau-Kle ne synd ome: a clinical and EEG s udy o i e cases. Epilepsia 31, 756-767. Holmes, G.L., Lenck-San ini, P-P., 2006. Role o he epilep i o m abno mali ies in cogni i e impai men . Epilepsy & Beha io 8, 504-515. ILAE Commission on Classi ica ion and Te minology: P oposal o e ised classi ica ion o epilepsies and epilep ic synd omes. 1989 Epilepsia 30: 389- 399. Jaseja, H., 2006. T ea men o in e ic al epilep i o m discha ges in ce eb al palsy pa ien s wi hou epilepsy: hope o a be e ou come in p ognosis. Clin. Neu ol. Neu osu g. 109, 221-224. Kas eleijn-Nols T eni é, D.G.A., Siebelink, B.M., Be ends, S.G.C., an S ien, J.W., Meina di, H., 1990. La e alized e ec o subclinical epilep i o m EEG discha ges on scholas ic pe o mance in child en. Epilepsia 31, 740-746. Lan z, G., Michel, C.M., Pascual-Ma qui, R.D., Spinelly, L., Seeck, M., Se i, S., Landis, T., Rosen, I., 1997. Ex ac anial localiza ion o in ac anial in e ic al epilep i o m ac i i y using LORETA (low esolu ion elec omagne ic omog aphy). Elec oenceph. Clin. Neu ophysiol. 102, 414-422. Massa, R., de Sain -Ma in, A., Ca cangiu, R., Rudol , G., Seegmulle , C., Klei z, C., Me z-Lu z, M.N., Hi sch, E., Ma escaux, C., 2001. EEG c i e ia p edic i e o complica ed e olu ion in idiopa hic olandic epilepsy. Neu ology 57, 1071-1079. McCo mick, D.A., Con e as, D., 2001. On he cellula and ne wo k bases o epilep ic seizu es. Annu. Re . Physiol. 63, 815-46. Me z-Lu z, M.N., Klei z, C., Desain ma in, A., Massa, R., Hi sch, E., Ma escaux, C., 1999. Cogni i e de elopmen in benign ocal epilepsies o childhood. De . Neu osci. 21, 182-190. Michel, C.M., Mu ay, M.M., Lan z, G., Gonzales, S., Spinelli, L., G a e de Pe al a, R., 2004. EEG sou ce imaging. Clin. Neu ophysiol. 115, 2195-2222.