PresurgicalAssessmentoftheEpilepsieswithClinicalNeurophysiologyandFunctionalImaging HandbookofClinicalNeurophysiology,Vol.3 FelixRosenowandHansO.Lüders(Eds.) ©2004ElsevierB.V.Allrightsreserved 3 CHAPTER1.1 Overview Felix Rosenowa,∗ and Hans O. Lu¨dersb aInterdisciplinaryEpilepsyCenter,DepartmentofNeurology,Philips-UniversityMarburg,D-35033Marburg,Germany bDepartmentofNeurology,S90,ClevelandClinicFoundation,Cleveland,OH44195,USA 1. Aimsandconceptsinsurgeryforepilepsy clinical neurophysiological assessment of surgically treatableepilepsies. The objective of resective epilepsy surgery is the complete resection or complete disconnection of the 2. Theconceptofcorticalzones epileptogeniczone,whichisdefinedastheareaofcor- texindispensableforthegenerationofclinicalseizures 2.1. Symptomatogeniczone (CarrenoandLu¨ders,2000).Ideally,thisaimshouldbe accomplishedwithoutdamagingthe“eloquent”cortex. The symptomatogenic zone is the area of cortex Currently,theepileptogeniczonecannotbemeasured which, when activated by an epileptiform discharge, directly. Therefore, a variety of diagnostic tools such produces the ictal symptomatology. It is defined by as analysis of seizure semiology, neurophysiological carefulanalysisoftheictalsymptomatologyeitherwith techniques, functional testing as well as structural a thorough seizure history or analysis of ictal video and functional neuroimaging, are used to indirectly recordings.Theprecisionwithwhichwecandefinethe definethelocationandboundariesoftheepileptogenic locationofthesymptomatogeniczonedependsonthe zone. These diagnostic methods define different specific ictal symptomatology. For example, a highly corticalzones(symptomatogeniczone,irritativezone, localized somatosensory aura such as paresthesias of ictal-onset zone, stimulation-induced seizure zone, oneortwofingersatthebeginningofaseizureclearly functional-deficit zone, and the epileptogenic lesion) localizesthesymptomatogeniczonetothecorrespond- whichareallamoreorlesspreciseindexoftheloca- ing primary sensory area. However, a poorly defined tionandextentoftheepileptogeniczone(Lu¨dersand bodysensationhaslittlelocalizingorlateralizingvalue. Awad,1992;RosenowandLu¨ders,2001).Ourability There are many other ictal signs or symptoms whose to precisely define these zones is essential to best localizing/lateralizingvaluefallsinbetweenthesetwo appreciatethetopographyoftheepileptogeniczone. extremes(seealsoChapter2.14). This volume is divided into four sections. The It is important to stress here that there is often no first section describes conceptually the different overlap between the symptomatogenic zone and the corticalzoneswhichbuildthestructureofthevolume. epileptogenic zone. The presence of ictal symptoma- Section2dealsindetailwiththedifferentneurophys- tologymaybecausedbygenerationoftheseizurefrom iological, functional imaging and other techniques a zone of eloquent cortex. However, in many cases, used to define these zones in candidates for epilepsy theictalsymptomatologyisproducedbyspreadofan surgery. In Section 3, the relative value and order epileptiformdischargefromanepileptogeniczonelo- of application of these techniques in the presurgical cated in “silent” cortex to a distant area of eloquent evaluation of epilepsy syndromes of different etiolo- cortex. giesarediscussed.Finally,Section4reportsindetail Electrical stimulation studies have demonstrated on ongoing clinical and basic research studies in the thatonly“strong”enoughstimulusparametersapplied toeloquentcortexwillelicitsymptoms.Inotherwords, only trains of stimuli that are long enough, have ap- ∗Correspondenceto:FelixRosenow,MD. propriatefrequency,andareofsufficientdurationand E-mailaddress:[email protected] intensitywillelicitsymptomsorsigns.Itisreasonable Tel.:+49-6421-286-5200;fax:+49-6421-286-5228. to expect also that only epileptiform discharges that 4 F.ROSENOWANDH.O.LU¨DERS Table1 Descriptionsofzonesandlesionsofthecortex(adaptedfromLu¨dersandAwad,1992) 1.Epileptogeniczone Regionofcortexthatcangenerateepilepticseizures;bydefinition,totalremovalor disconnectionoftheepileptogeniczoneisnecessaryandsufficientforseizurefreedom 2.Irritativezone RegionofcortexthatgeneratesinterictalepileptiformdischargesintheEEGorMEG 3.Seizure-onsetzone Regionfromwhichtheclinicalseizuresoriginate 3.1.Stimulation-inducedseizurezone Regionofcortexwhoseelectricalstimulationevokesseizures 4.Epileptogeniclesion Structurallesionthatiscausallyrelatedtotheepilepsy 5.Ictalsymptomatogeniczone Regionofcortexthatgeneratestheinitialseizuresymptomatology 6.Functional-deficitzone Regionofcortexthatintheinterictalperiodisfunctionallyabnormalasindicatedby neurologicalexamination,neuropsychologicaltesting,functionalimaging,or nonepileptiformEEGorMEGabnormalities 7.Eloquentcortex Regionofcortexthatisindispensableforcertaincorticalfunctions fulfillthesecriteriawillelicitsymptomsorsigns.This 2.3. Seizure-onsetzone explainsthefrequentobservationthatanepileptiform The seizure-onset zone is the area of cortex from discharge is recorded from a symptomatogenic zone which clinical seizures are currently generated. Like without producing the corresponding symptoms. All the irritative zone, it is most commonly localized by the limitations outlined above have to be considered eitherscalporinvasiveEEGtechniques(seeChapters whentryingtodefinetheepileptogeniczonebycareful 2.6–2.9).Unliketheirritativezone,however,theloca- analysisofictalsymptomatology(Table1). tionoftheseizure-onsetzonecanalsobedetermined byictalsinglephotonemissioncomputedtomography 2.2. Irritativezone (SPECT;seeChapter2.10).Theseizure-onsetzoneis Theirritativezoneisdefinedastheareaofcortical usuallythatportionoftheirritativezonethatgenerates tissue that generates interictal electrographic spikes. spikes capable of producing afterdischarges. These Interictalspikescanbeconsideredas“mini-seizures”. consistofrepetitivespikesthathaveenough“strength” Isolatedspikesthatareofsufficient“strength”andare to produce clinical ictal symptoms when generated generatedwithinaneloquentcorticalareaactuallycan in eloquent cortex or when spreading into eloquent give rise to clinical symptoms. Typical examples are cortex. For many years, it was thought that precise myoclonicjerksthatcanbeseeninpatientswithspikes definition of the seizure-onset zone should provide intheprimarymotorcortex.Ingeneral,however,iso- an accurate definition of the epileptogenic zone, but lated,independentspikeswillnotgenerateanyclinical thisisnottrue.Actually,thereisnomethodologythat symptoms regardless of whether they are located in permits precise definition of the location or extent silent or eloquent cortex. In general, only runs of of the seizure-onset zone. Scalp electrodes give us epileptiform discharges (afterdischarges) of sufficient an excellent overview of the electrical activity of the “strength”inducesymptomswhentheyinvadeasymp- brainandfrequentlysuggestthesideandapproximate tomatogeniczone.Theirritativezoneusuallyoverlaps locationoftheseizureonset.However,scalpEEGhasa with the epileptogenic zone but is more extended. relativelylowsensitivityfortheprecisedetectionofthe However,inmanycases,spikescanbegeneratedfrom seizureonsetbecausesurfaceelectrodesarelocatedat areasdistantandevencontralateraltotheepileptogenic arelativelylargedistancefromthecortexandaresepa- zone (for examples, see Chapters 3.1 and 3.2). The ratedfromthebrainbyaseriesofbarriersthatinterfere irritativezoneisdefinedbyscalporinvasiveelectroen- significantlywiththetransmissionoftheelectricalsig- cephalography (EEG) or magnetoencephalography nals. Usually, afterdischarges generated locally at the (MEG;Chapters2.1–2.4).Automaticspikedetection, actualseizure-onsetzonearetoosmalltobedetected discussed in Chapter 2.5, can help to reduce the by scalp electrodes, and therefore, scalp electrodes review time needed to define the complete irritative areonlycapableofdetectingaseizuredischargeafter zone but carries the potential risk of introducing it has spread considerably (see also Chapter 2.6). aselectionbias. Invasive cortical surface electrodes, however, also OVERVIEW 5 havesignificantlimitations.TheyrecordEEGactivity of the SISZ does not allow prediction of seizure out- from only an extremely limited region of the brain. come(Schulzetal.,1997).Therefore,thedelineation By eliminating distance and insulating barriers, each oftheSISZplaysarelevantroleinthedefinitionofthe invasiveelectroderecordstheEEGactivitygenerated epileptogeniczoneonlyinaminorityofcandidatesfor in the limited cortical area covered by that electrode. epilepsysurgery. Invasiveelectrodesare,therefore,extremelysensitive for the detection of afterdischarges but will only be 2.4. Epileptogeniclesion abletoaccuratelydefinetheseizureoriginiftheyare implanteddirectlyintheseizure-onsetzone.Unfortu- This is a lesion visible by neuroimaging that is the nately, since only a limited number of electrodes can causeoftheepilepticseizures.Thebestwaytoidentify be implanted, invasive electrodes are usually located theepileptogeniclesionisbyhigh-resolutionmagnetic intheperipheryoftheseizure-onsetzoneoronlycover resonanceimaging(MRI;seeChapter2.15).However, aportionofthiszone(seealsoChapters2.7–2.9). some lesions identified by neuroimaging may not be The second problem to consider is that the extent epileptogenicandbeunrelatedtotheclinicalseizures. of the seizure-onset zone may not correspond with Forthisreason,evenwhenweseealesionintheMRI, theextentoftheepileptogeniczone(thecorticalarea we still have to use other methods to verify (usually from which clinical seizures may arise). In other by video/EEG monitoring and/or seizure semiology) words, the epileptogenic zone may be either more or thatthelesionvisualizedbyneuroimagingisindeedre- less extensive than the seizure-onset zone. When the sponsibleforthepatient’sseizures.Arelatedproblem epileptogenic zone is smaller than the seizure-onset isthedefinitionofepileptogenicityincaseswithdual zone, partial resection of the seizure-onset zone ormultiplepathology.Hereagain,additionaltestingis may lead to seizure freedom because the remaining necessary to define which of the lesions are epilepto- seizure-onsetzoneisnolongercapableofgenerating genic.Inthosecasesinwhichtwoormorelesionsare seizures. Conversely, when the epileptogenic zone is placed in close spatial proximity, the problem of at- greaterthantheseizure-onsetzone,wehaveasituation tributingepileptogenicitytoonelesionoranothercan inwhicheventotalresectionoftheseizure-onsetzone frequently only be resolved with the use of invasive will not result in seizure freedom. This phenomenon EEGtechnology(seeChapters2.7and2.8). occurs when a patient has seizure-onset zones of The spatial relationship of the epileptogenic zone different thresholds within a single epileptogenic withtheepileptogeniclesionissimilartoitsrelation- zone.Theseizure-onsetzoneoflowestthresholdwill ship with the seizure-onset zone that was discussed generate all the usual seizures and is, therefore, the above. It has been thought that complete resection of only one we can directly measure before surgery. theepileptogeniclesionvisualizedbyneuroimagingis Once this zone has been resected, however, another necessarytoobtainseizurefreedom.Thisisnotalways seizure-onset zone of higher threshold may become true, however, as evidenced by cases in which only clinically evident. The cortical area of interest now partial lesion resection (often unavoidable because of becomes a “potential epileptogenic zone” containing partial overlap between the epileptogenic lesion and one or more seizure-onset zones. There is no way to eloquent cortex) results in complete seizure freedom. predictwithcurrenttechnologywhetheranyadditional However, often seizures persist in spite of complete higherthresholdseizure-onsetzonesexist. resectionofthelesionvisibleonMRI.Thisisoftenthe caseinpatientswithcorticaldysplasiaorposttraumatic 2.3.1. Stimulation-inducedseizurezone epilepsy (see also Chapters 3.4 and 3.5). There are Duringelectricalstimulationofthecortex,afterdis- twopossibleexplanationsforthisphenomenon.Many charges and seizures, including the habitual seizures lesions are not intrinsically epileptogenic but induce of the patient, can be elicited. The area from which seizures by generating reactions in the surrounding seizures(includingauras)areevokedisthestimulation- braintissuewithwhichtheyareincontact.Anotherex- inducedseizurezone(SISZ).Ithasbeensuggestedthat planationaddressesthesensitivityofMRIindetecting cortex from which seizures can be elicited has facil- the complete lesion. Brain tissue adjacent to lesions itated pathways to a specific symptomatogenic zone visiblebyneuroimagingmayconsistoflesionaltissue andis,therefore,likelytobepartoftheepileptogenic of a lesser pathologic severity. This tissue, while fre- zone(seeChapter2.13).However,completeresection quently of high epileptogenicity, may be invisible on 6 F.ROSENOWANDH.O.LU¨DERS MRI.Thisoccurrsfrequentlyinpatientswithcortical pocampalsclerosis.PETstudiesoftenrevealextensive dysplasia in which only the “tip of the iceberg” is hypometabolic regions outside the mesial temporal visible in MRI. Failure to resect these MRI-invisible structures (see Chapter 2.16) even when we know lesionscanleadtopersistenceofseizuresafterepilepsy that the epileptogenic zone is usually limited to the surgery. This is the most likely explanation for the mesialtemporalregion,andseizurefreedomisusually relativelyhighfrequencyofsurgicalfailureinpatients achievedbyaselectiveamygdalo-hippocampectomy. withneocorticaldysplasia(seealsoChapter3.4). In spite of all these limitations, a good correlation Theonlywaywecantrytopredictthepresenceof ofthefunctional-deficitzonewiththeotherzonesde- a perilesional epileptogenic rim is by understanding finedabovegivesusadditionalsupportiveinformation better the pathological nature of the MRI-visible regardingthelateralizationorlocationoftheepilepto- lesion. We know, for example, that well-delineated geniczone.However,cleardiscrepanciesbetweenthe braintumorsandcavernousangiomastendtoproduce resultsobtainedinthetestsusedtodefinethedifferent epileptogenicity only in the MRI-visible lesion and cortical zones make accurate definition of the epilep- its immediate surrounding. Therefore, lesionectomy togeniczoneverydifficultandarefrequentlyareason is usually successful in these cases. However, as to request more sensitive, specific tests like invasive mentioned above, cases with cortical dysplasia monitoring. or posttraumatic epilepsy typically require more extensiveresectionsforasuccessfuloutcome. 2.6. Epileptogeniczone 2.5. Functional-deficitzone The epileptogenic zone is the area of cortex that is indispensableforthegenerationofepilepticseizures. Thisisdefinedastheareaofcortexthatisfunction- Itmayincludeanactualepileptogeniczonethatisthe allyabnormalintheinterictalperiod.Thisdysfunction cortical area generating seizures before surgery (see may be a direct result of the destructive effect of the above:equivalenttoorsmallerthantheactualseizure- lesionormaybefunctionallymediated,i.e.abnormal onsetzone)anda“potentialepileptogeniczone”,which neuronal transmission that may affect brain function is an area of cortex that may generate seizures after either locally or at a considerable distance from the the presurgical seizure-onset zone has been resected. epileptogenic tissue. There are a variety of methods Thereisnodiagnosticmodalitythatcanbeusedtodi- that can be used to measure the functional-deficit rectly measure the entire epileptogenic zone. This is zone. Examples include neurological examination, becausewecannotexcludethepossibleexistenceofa neuropsychological testing, EEG, positron emission “potentialepileptogeniczone”thatwouldonlybecome tomography (PET) scan, and interictal SPECT (see clinicallyapparentpostoperatively.Theepileptogenic Chapters2.16and2.17). zone,therefore,isatheoreticalconcept.Ifthepatientis What is the relationship between the functional- seizure-freeaftersurgery,weconcludethattheepilep- deficit zone defined by these different tests and the togenic zone must have been included in the surgical location of the epileptogenic zone? Unfortunately, resection. this relationship is very complex and difficult to Since we cannot measure the epileptogenic zone establish even in individual cases. This is related to directly, we must infer its location indirectly by the fact that these tests measure parameters such as definingtheotherzonesdiscussedabove.Whenallof global brain function (general neurological exam) these data are concordant, the determination is easy. or brain physiology (local glucose metabolism or Unfortunately, these cases are rare. In most cases, blood perfusion), which are not directly related to there is some degree of discrepancy between the epileptogenesis. Some of these changes may be the differentzones.Attemptsshouldthenbemadetofind resultofanonepileptogeniclesionandmayoccurata aplausibleexplanationforthesediscrepancies,taking considerabledistancefromtheprimaryseizurefocus. intoconsiderationthebasicprinciplesoutlinedabove. Forthisreason,thevalueofdefiningpresurgicallythe It is difficult to accurately define the epileptogenic functional-deficit zone is relatively limited compared zonewhennoadequateexplanationforthediscrepant to measurement of the other zones discussed above locationand/orextentofthedifferentzonesisfound. which are more directly related to the seizures. A Inthesecases,surgeryshouldbedeferredwhilemore typicalexampleofthisisseeninpatientswithpurehip- precise testing like invasive video/EEG is performed. OVERVIEW 7 As discussed earlier, however, invasive recordings thathopefullywillpermitmoredirectdefinitionofthe are limited by their extent of cortical coverage. They epileptogeniczone.Itisverylikelythatthesedevelop- shouldonlybeusedinthosecasesinwhichthereisa mentswillbeinfunctionalneuroimaging.Allcurrently clearhypothesisregardingthelocationofseizureonset widely available functional neuroimaging techniques and in which a specific question must be answered (mainlyFDG-PETandinterictalSPECT)measureonly (seealsoChapters2.7–2.9). nonspecificbrainphysiology,likeregionalmetabolism or blood flow. However, further developments may 2.7. “Eloquent”cortex make it possible to directly image the distribution of neurotransmitters and receptors involved in the Epilepsy surgery is usually an elective procedure. pathogenesisofepilepsy.Thismaynotonlyallowusto While the complete resection of the epileptogenic definedifferenttypesofepileptogeniclesionsbasedon zone is of paramount importance, this aim is limited neurotransmitterandreceptorphysiology,butalsogive by one major restriction: the sparing of eloquent us a measurement of “potential epileptogenic zones” cortex to avoid new, unacceptable deficits for the thatarecurrentlyundetectablepreoperatively.Inaddi- patient. Eloquent cortex is cortex necessary for a tion,refinementsofthecurrentlyavailablediagnostic given function. Methods commonly used to detect techniques may increase the accuracy with which we such functions include electrical stimulation of the define the different zones giving us some additional cortex, evoked potentials, MEG, functional magnetic power,eveniftheydonotsolvesomeoftheessential resonanceimaging(fMRI),and,toalesserextent,PET theoretical limitations discussed before. In Section 4, (see Chapters 2.17–2.24). A possible loss of function several of the newer diagnostic techniques are pre- as a consequence of the epilepsy surgery should at sented,andtheircurrentandfuturevalueinthepresur- least be predictable and discussed with the patient gical diagnosis of patients with epilepsy is discussed prior to the procedure. Even today, invasive methods indetail. such as subdural electrodes are still frequently used whenthelocationandextentoftheseizure-onsetzone References and of eloquent cortex are to be defined. Hopefully, thesetechniqueswillhavetobeusedlessfrequentlyin Carreno, M and Lu¨ders, HO (2000) General principles of the future when novel neurophysiology and imaging presurgical evaluation. In: HO Lu¨ders and YG Comair techniquesdiscussedinSection4ofthisvolumewill (Eds.),EpilepsySurgery.LippincottWilliams&Wilkins, have become more widely available as reliable tools Philadelphia,pp.185–199. inthepresurgicaldiagnosisofepilepsy. Lu¨ders,HOandAwad,I(1992)Conceptualconsiderations. In:HOLu¨ders(Ed.),EpilepsySurgery.RavenPress,New York,pp.51–62. 3. Futureperspectivesinthedefinitionofthe Rosenow,FandLu¨ders,H(2001)Presurgicalevaluationof epileptogeniczone epilepsy.Brain,124:1683–1700. Schulz, R, Lu¨ders, HO, Tuxhorn, I, Ebner, A, Holthausen, Aswementionedabove,theepileptogeniczoneisa H, Hoppe, M, Noachtar, S, Pannek, H, May, T and theoreticalconcept.Noneoftheavailabletestspermits Wolf, P (1997) Localization of epileptic auras induced directmeasurementoftheepileptogeniczone.Inthefu- on stimulation by subdural electrodes. Epilepsia, 38: ture,wewillhavetolookfornewdiagnostictechniques 1321–1329. PresurgicalAssessmentoftheEpilepsieswithClinicalNeurophysiologyandFunctionalImaging HandbookofClinicalNeurophysiology,Vol.3 FelixRosenowandHansO.Lüders(Eds.) ©2004ElsevierB.V.Allrightsreserved 11 CHAPTER2.1 Noninvasive EEG in the definition of the irritative zone ∗ Hajo M. Hamer and Niki Katsarou DepartmentofNeurology,UniversityofMarburg,Rudolf-Bultmann-Str.8,D-35033Marburg,Germany 1. Generalconcepts ing that a strong after-inhibition produced by IED protects against the occurrence of ictal discharges by The main goal of the presurgical evaluation in maintainingalowlevelofexcitationinageneralcon- patients with medically intractable epilepsy is the dition of hyperexcitability (Binnie and Stefan, 1999; identificationofthecorticalareacapableofgenerating DeCurtisandAvanzini,2001).Thisfinding,however, seizures, and whose removal or disconnection will remainscontroversial(GotmanandKoffler,1989). result in seizure freedom. This area is called the “epileptogenic zone” (Carreno and Lu¨ders, 2001). 1.1. Limitsofscalpandsphenoidalelectrodes The irritative zone is the area of cortex capable of comparedtointracranialelectrodes generating interictal epileptiform discharges (IED) Several square centimeters of synchronously in the EEG. The irritative zone does not coincide discharging cortex (around 6cm2) are necessary for but frequently overlaps with the epileptogenic zone. detectionbyoverlyingscalpelectrodes(Cooperetal., Extensive experience with routine extracranial EEG 1965;FernandezTorreetal.,1999b).Cerebralactivity shows that in general, the location of IED is a good is attenuated by the impedance of the cerebrospinal indicator of the area of cortex from which seizures fluid, meninges, skull, and scalp. Therefore, scalp areoriginating(BinnieandStefan,1999;Carrenoand EEGfailstorecordagreatnumberofIEDrecordedby Lu¨ders, 2001). The methodology of many studies on depthorsubduralelectrodes(Fig.1).Interictalactivity this topic, however, is biased towards patients with arising from deep or midline structures is usually not medically refractory epilepsies (Foldvary, 2000; So, reflected in scalp EEG (Carreno and Lu¨ders, 2001). 2001). The term “interictal epileptiform discharges” The distribution of IED on the scalp depends on the includesspikesandsharpwaves,whicharewaveforms conductive properties of the surrounding tissue, the distinguishablefrombackgroundactivity.Bothspikes spatialcharacteristicsofthegenerator,andthespatial with a duration of less than 80ms and sharp waves resolution of the surface EEG. Consequently, the with a duration of 80–200ms have a pointed peak distributionofinterictalepilepticdischargescanfailto and are usually of negative polarity because they are localizeorevenmislocalizetheregionorhemisphere generated by depolarization of vertically oriented of seizure origin (Foldvary, 2000). Closely spaced neurons (Celesia and Chen, 1976; Pedley and Traub, scalp electrodes can improve the yield of spike 1997; Foldvary, 2000). Positive IED can be seen in detection and localization over the standard 10–20 electrocorticographyandinfrequentlyinpatientswith system(Morris,IIIetal.,1986;Hameretal.,1999). a breach rhythm, head trauma, or cortical dysplasia It remains controversial whether or not sphenoidal (Daly and Pedley, 1990; Foldvary, 2000). Spikes and electrodes increase the detection sensitivity of IEDs. sharpwavesconveyanincreasedriskofepilepsyand Whereasseveralstudiesfoundthatinaminorityofpa- mustbedifferentiatedfrombenignvariantsresembling tients, especially those with temporal lobe epilepsies, IED(BinnieandStefan,1999;Foldvary,2000).Some sphenoidalelectrodescanrecordepileptiformactivity reports suggest that spike suppression, and not an in- whichdoesnotappearinanteriortemporalelectrodes creaseinIEDfrequency,heraldsseizures,thuspropos- (Fig. 2; Marks et al., 1992; Tuunainen et al., 1994; Kanner and Jones, 1997; Pacia et al., 1998; Provini ∗Correspondenceto:HajoM.Hamer. et al., 1999), others could not verify this observation E-mailaddress:[email protected] (FernandezTorreetal.,1999a);seealsoChapter2.6. Tel.:+49-6421-286-5200;fax:+49-6421-286-5208. Spikes with maximal amplitude at the sphenoidal 12 H.M.HAMERANDN.KATSAROU Fig.1. Invasivevideo-EEGmonitoringwithbilateralforamenovaleelectrodesofa38-year-oldpatientwithbitemporalepilepsy. Inaddition,scalpelectrodeswereattachedaccordingtotheInternational10–20system,andsphenoidalelectrodeswerebilaterally inserted.NotetheIEDrecordedbythefirsttwocontactsoftheleftforamenovaleelectrodewhilethescalpandsphenoidal electrodesfailtodetecttheseIED.FOr1,FOr2,FOr3:threecontactsoftherightforamenovaleelectrode;FOl1,FOl2,FOl3: threecontactsoftheleftforamenovaleelectrode;Sp1:leftsphenoidalelectrode;Sp2:rightsphenoidalelectrode. electrodescanarisefromthemesialorlateraltemporal healthy young men (Gregory et al., 1993) to 12% in lobe or from an orbitofrontal focus (Marks et al., a study including all age groups and patients with 1992). Sphenoidal electrodes placed close to the progressive cerebral disorders (Sam and So, 2001). foramen ovale can register higher amplitudes of IED Specificityisprobablylowerandsensitivityhigherin thanscalpelectrodes(Morris,IIIetal.,1989;Kanner children than in adults, but reliable estimates are not et al., 1995; Foldvary, 2000). However, sphenoidal available(BinnieandStefan,1999).Thefrequencyof electrodesstillfailtodetectalargeproportionofIED IED in elderly patients with epilepsy is substantially recordedbyforamenovaleelectrodes(FernandezTorre lower than that reported in epileptic populations as a etal.,1999a)(Fig.1). whole(MarsanandZivin,1970;DruryandBeydoun, 1998). The majority of reports could not establish a correlation between the levels of anticonvulsants 1.2. YieldofscalpEEG and the frequency of interictal spiking, although this The first EEG will uncover IED in about 30–50% issueremainscontroversial(MarsanandZivin,1970; of the patients with epilepsy, but the yield increases Gotman and Marciani, 1985; Duncan, 1987; Gotman to 60–90% by the fourth EEG (Marsan and Zivin, and Koffler, 1989; Sundaram et al., 1990; Eriksson 1970; Salinsky et al., 1987; Binnie and Stefan, 1999; etal.,2001).Overall,around10%ofthepatientswith Blume, 2001a). This contrasts with the frequency of epilepsy would be expected to show no IED in scalp IED in nonepileptic patients ranging from 0.5% in EEGduringwakefulnessorsleepinspiteofprolonged NONINVASIVEEEGINTHEDEFINITIONOFTHEIRRITATIVEZONE 13 Fig.2. TwoexamplesofIEDofa39-year-oldpatientwithrightmesialtemporalepilepsyduetohippocampalsclerosis.Scalp electrodes were attached according to the International 10–20 system, and additional sphenoidal electrodes were bilaterally inserted(Sp2:rightsphenoidalelectrode).Thefirstfourchannelsrepresentalongitudinalbipolarmontageofscalpelectrodes overtherighttemporalarea.Inthefollowingfivechannels,therightsphenoidalelectrodeisincludedbetweenF8andT8ina bipolarfashion.Notethesharpwavesrecordedbythesphenoidalelectrodeandmissedbythescalpelectrodes. or repeated recordings (Chung et al., 1991; Adachi (Sammaritano et al., 1991; Malow et al., 1997, 1998; etal.,1998;BinnieandStefan,1999).Theyieldfrom MendezandRadtke,2001;CarrenoandLu¨ders,2001). a single EEG is substantially increased in patients Increased neuronal synchronization within thalam- investigated within 1 or 2 days after a seizure, and is ocortical projection neurons during non-REM sleep greaterinpatientswithmonthlyseizuresthaninthose may contribute to the activation process in epileptic who had been seizure-free for a year (Marsan and cortex (Adachi et al., 1998). The predictive accuracy Zivin,1970;Sundarametal.,1990;Kingetal.,1998). inlateralizationofepileptogenesisisimprovedduring Thedurationofrecordingmayalsoaffectthedetection non-REM sleep in patients who showed bilateral rateofinterictalspiking(GotmanandKoffler,1989). or no discharges in wakefulness because unilateral Moreover, the yield can be significantly increased, discharges arising de novo in sleep were always and new abnormalities found, if the EEG includes correctlylateralizing(Adachietal.,1998).Inpatients sleeprecordings(Kingetal.,1998;BinnieandStefan, who had unilateral discharges in the awake state, 1999). IED are seen more commonly during sleep, whetheripsilateralorcontralateraltotheepileptogenic with the greatest activation during non-REM sleep zone, the findings were generally unchanged during 14 H.M.HAMERANDN.KATSAROU sleep. Sleep may also alter IED morphology and Chapter2.4).Severalstudiesfoundanassociationbe- distribution (Mendez and Radtke, 2001). Spikes can tweenpostresectionIEDinthistestandalessfavorable be more widespread during non-REM sleep than outcome (So et al., 1989; Fiol et al., 1991; McBride duringwakefulnessorREMsleep,suggestingthatthe et al., 1991; Salanova et al., 1996; Wennberg et al., localizationofinterictaldischargesinREMsleepmay 1998),whileothersdidnot(Godoyetal.,1992;Shih beabetterindicatoroftheepileptogeniczonethanin etal.,1994;Tuunainenetal.,1994;Tranetal.,1995; non-REM sleep (Sammaritano et al., 1991). Hyper- Radhakrishnan et al., 1998; Ficker et al., 1999; ventilation or photic stimulation rarely activates IED McIntosh et al., 2001). The lack of agreement of in patients with focal epilepsies (Celesia et al., 1975; these studies may be due to differences in the patient Miley and Forster, 1977; Kasteleijn-Nolst Trenite, populations (e.g. lesional versus nonlesional cases), 1989;Shiraishietal.,2001;Kleinetal.,2003). recording techniques, and anesthetic agents used during surgery. Even in the studies confirming the association of postresection IED and seizure contin- 1.3. SpecificityofinterictalscalpEEG uation, the percentage of patients with postresection In general, the frequency, repetition rate, morpho- persistenceofIEDbutstillafavorableoutcomevaried logiccharacteristics,andstatedependenceofinterictal from 26 to 47% (So et al., 1989; Fiol et al., 1991; epileptiform activity cannot be used to predict the Salanova et al., 1996), which can make it difficult to etiology of the disorder (Foldvary, 2000). However, estimatetheprognosisinindividualcases. rhythmic spiking on a slow background activity and not associated with behavioral changes has been 2. Temporallobeepilepsy(TLE) foundtobecharacteristicforfocalcorticaldysplastic lesions(Palminietal.,1995;Gambardellaetal.,1996; Asmentionedabove,severalstudieshaveindicated Rosenowetal.,1998).Tumorstendedtocausewider that the majority of mesial and neocortical temporal distributedIEDascomparedtodevelopmentalabnor- IEDcannotberecordedbysurfaceelectrodes(Cooper malitiesorhippocampalsclerosis(Aykut-Bingoletal., et al., 1965; Alarcon et al., 1994; So, 2001; see 1998; Hamer et al., 1999; Kameyama et al., 2001), also Chapter 3.1). The IED which are recorded in which was also true for seizure patterns (Bo¨sebeck the extracranial EEG, however, may show a better et al., 2002). This may be caused by local neuronal correlationwiththeseizureoriginascomparedtoIED injury,edema,ischemiaorotherelectricaland/orbio- recorded by depth electrodes (Blume et al., 1993). chemicaleffectsofthestructurallesiononsusceptible This suggests that widely synchronous interictal neighboring brain tissue or homologous contralateral spikes more likely rise from the epileptogenic area areas(Spenceretal.,1984;Awadetal.,1991). thanspikeswithsmallerfields. In addition to IED, temporal intermittent rhythmic 1.4. PrognosticrelevanceofpostoperativeIED delta activity (TIRDA) was also strongly correlated withaclinicaldiagnosisofTLEandcouldberecorded In both temporal and extratemporal epilepsy, inalargenumberofpatientswithmedicallyrefractory absenceofIEDinthe6-monthor1-yearpostoperative TLE (Fig. 3; Reiher et al., 1989; Blume et al., 1993; scalp EEG was associated with good postoperative Normand et al., 1995; Gambardella et al., 1996). outcome (Godoy et al., 1992; Patrick et al., 1995; However, temporal intermittent polymorphic delta Morris et al., 1998; Aronica et al., 2001). The prog- activity was recorded in patients with temporal and nosticvalueofa3-monthpostoperativeEEGremains extratemporalepilepsies(Geyeretal.,1999). controversial(Cascinoetal.,1992;Kirkpatricketal., 1993; Tuunainen et al., 1994; Radhakrishnan et al., 2.1. DistributionofIEDinTLE 1998). The presence of IED in the early extracranial postoperative EEG (within 1 or 2 weeks) was not In TLE, IED tend to produce a stereotyped pattern foundtobeofprognosticvalueinmostofthestudies onthescalpwiththehighestamplitudesattheanterior (Soetal.,1989;Salanovaetal.,1992;Radhakrishnan temporal electrodes. This may be due to the location et al., 1998; Ficker et al., 1999). There are conflict- of the neuronal generators within the temporal lobe ing results on the prognostic value of IED in the and anatomical characteristics of the brain coverings, postresection electrocorticography (ECoG; see also such as skull discontinuities (Fernandez Torre et al., NONINVASIVEEEGINTHEDEFINITIONOFTHEIRRITATIVEZONE 15 Fig.3. InterictalEEGofa21-year-oldpatientwithlefttemporallobeepilepsyduetohippocampalsclerosis.Scalpelectrodes areattachedaccordingtotheInternational10–20system.NoteTIRDAinthelefttemporalregion. 1999b). Children may have a more widespread irrita- and can reflect bilateral damage, dysfunction at a tive zone, which may be caused by a high frequency distance, or secondary epileptogenesis (Margerison of dual pathology (Mohamed et al., 2001). Most and Corsellis, 1966; Morrell, 1985; Sammaritano (Bengzon et al., 1968; Dodrill et al., 1986; Barry etal.,1987). etal.,1992;Keoganetal.,1992;Salanovaetal.,1996; InextracranialEEGrecordings,interictaldischarges McIntosh et al., 2001), but not all, studies (Chung can have a more reliable lateralizing value than ictal etal.,1991;Cascinoetal.,1996)confirmedthefinding changesintemporallobeepilepsy(Sammaritanoetal., that IED confined to the anterior temporal region 1987), and they almost always predict seizure origin are predictive of a good postoperative outcome in inTLEandgoodpostoperativeoutcomeiftheyareex- TLEsurgery. clusiveorclearlypreponderantonthesideofsurgery (Blume et al., 1993; Chee et al., 1993; Hufnagel et al., 1994; Salanova et al., 1994, 1996; Steinhoff 2.2. LateralizingvalueofIED etal.,1995;Holmesetal.,1996;Pataraiaetal.,1998; TheincidenceofbilateralIEDinTLEisestimated Radhakrishnanetal.,1998;Serlesetal.,1998;Malow to be between 20 and 44% and may be more when et al., 1999; Blume, 2001a; McIntosh et al., 2001; investigated with invasive EEG (So, 2001). The So, 2001). In a study of 59 candidates for temporal probability of bilateral IED was positively correlated lobectomy, 92% of patents with >90% lateralization with the duration of the EEG monitoring (Cascino hadagoodsurgicaloutcome,whereasonly50%with et al., 1996; Ergene et al., 2000). Bitemporal IED <90%hadafavorableoutcome(Chungetal.,1991). increase the likelihood that seizures are arising In addition to IED, lateralization of delta activity independendlyfrombothsides(Steinhoffetal.,1995) also indicates the side of temporal seizure origin in