sign in sign up
<<
Home , Electroencephalography

Current trends in electroencephalography Warren T. Blume

Several recent articles re-emphasize the value of clinical Introduction : in localizing , predicting Two current trends will continue to enhance and clarify effectiveness of surgery, and disclosing a mechanism of the role of electroencephalography (EEG) in clinical benign of childhood. neurophysiology. As EEG technology and statistical A review of the role of EEG in the diagnosis of epilepsy indicated analysis become more sophisticated, the clinical rele- that epileptiform activity will appear in 50% of initial awake vance of traditional EEG becomes more de®ned. More- recordings of adults with epilepsy and in 85% of subjects over, newer electrophysiological methods promise to undergoing two recordings. This contrasts with the appearance enhance further the value of traditional EEG. Some of spikes in only 4 of 1000 normal persons. Several studies recent ®ndings of these traditional and newer aspects of focused on the value of electroencephalography in extratemporal EEG form the subject of this concise overview. epilepsy: 62% of patients with neocortical epilepsy had at least one localizing ictal EEG; occipital and temporal neocortical Role of electroencephalography in the were localized in a greater proportion than frontal or diagnosis and management of epilepsy parietal attacks. Interictal spikes, if unifocal, always arose from Binnie and Stefan [1 .] reviewed the diagnostic sensitiv- the epileptogenic region in a study of their localizing ity and speci®city of EEG in epilepsy. Although value. Such congruence augured for better seizure control by interictal epileptiform activity is apparent in only focal resection in two studies reviewed herein. approximately 50% of single awake recordings in adults with epilepsy, this proportion rises to 80±85% if is Studies indicating the value of interictal temporal lobe spikes and included. Two awake recordings will demonstrate scalp-recorded seizures in lateralising a temporal seizure focus epileptiform activity eventually in 85% of individuals, are reviewed. One study found EEG to be slightly more reliable for and this rises to 92% of persons within four recordings. lateralization of temporal epileptogenesis than MRI. In patients with benign Rolandic seizures, enhanced motor These authors quoted studies that show that spikes evoked potentials (MEPs) were obtained from transcranial occur in four of 1000 normal persons. More relevant, magnetic stimulation when this was applied 50±80 msec after however, is the incidence of epileptiform activity in a electrical stimulation of the thumb whereas this interval inhibited patient population without a history of epilepsy. Ajmone the MEP in normal subjects. This suggests that afferent cutaneous Marsan and Zivin [2] found a 2±3% incidence, because input abnormally and synchronously activates a large population some systemic or nonepileptic cerebral illnesses may be of sensory ; such activation is subsequently transmitted to associated with spikes. the to produce the focal spikes in this condition. Binnie and Stefan [1 .] pointed out that the clinician is Finally, advances in non-invasive technology have redefined and often faced with an inadequate description of a clinical limited the need for invasive in children with event, and relies on the EEG to determine its nature. intractable seizure disorders. Curr Opin Neurol 14:193±197. # 2001 Evidence-based data must be interpreted in this light; Lippincott Williams & Wilkins. for example, Gilbert and Buncher [3] found EEG to have little value in predicting recurrence risk after a ®rst unprovoked seizure in childhood. However, EEG has several other uses in relationship to epilepsy, as outlined London Health Sciences Centre ± University Campus, The University of Western . Ontario, London, Ontario, Canada by Binnie and Stefan [1 ]. In addition to supporting the diagnosis of epilepsy, EEG may classify epileptic Correspondence to Dr. Warren T. Blume, London Health Sciences Centre ± University seizures into an epilepsy syndrome, it may monitor Campus, 339 Windermere Road, London, Ontario, Canada N6A 5A5 Tel: +1 519 663 3690; fax +1 519 663 3753 absences, it may recognize previously unidenti®ed or misclassi®ed attacks as seizures, and it may detect signs Current Opinion in 2001, 14:193±197 of antiepileptic drug intoxication. Abbreviations EEG electroencephalography Value of electroencephalography in MEG MEP motor- extratemporal epilepsy MRI magnetic resonance imaging Lee et al.[4.] studied the localizing value of ictal surface PET positron emission tomography EEG in neocortical epilepsy by analysing 394 ictal recordings from 86 neocortical epilepsy patients. All # 2001 Lippincott Williams & Wilkins 1350-7540 patients had neocortical epileptogenic foci, as con®rmed

193 194 Seizure disorders with presurgical invasive . Fifty-three of the some type of visual impairment. Occipital foci tend to patients (62%) had at least one localizing ictal EEG, and appear at early ages as indicated; relatively few patients 167 localizing EEGs were identi®ed in the 394 ictal in the group studied by Smith and Kellaway were older recordings (42%). An additional 10% of patients and 9% than 8 years, and spikes did not appear on subsequent of recordings gave lateralizing information whereas 49% recordings (age not indicated) in 38% of the patients of EEGs in 28% of the patients were unhelpful. Ictal studied by Libenson et al. EEG was of signi®cantly less value in simple partial seizures, a ®nding that has been noted by others. Electroencephalography and benign central temporal Occipital lobe seizures and temporal neocortical seizures (Rolandic) epilepsy of childhood could be localized in a greater proportion than could Kellaway [9] described the morphological features of frontal or parietal attacks with respect to both proportion `Rolandic' spikes and contrasted the topologically of patients and of seizures (Table 1). stationary and slightly propagating varieties. The pre- dominant negative spike may be preceded by a low- Holmes et al.[5.] also found the scalp EEG to re¯ect amplitude sharper surface positive spike, and is pre- ictal origin, but in a restricted percentage of patients. ceded by a trough followed by an after-coming slow They reviewed 126 patients with refractory extratempor- wave. al partial seizures who underwent . In 26 (21%), interictal spikes were unifocal and in all cases In an ingenious study, Manganotti and Zanette [10 ..] clinical seizures arose from the interictal spike region. utilized electrical stimulation of the ®ngers and motor- Follow up indicated that the most signi®cant predictor of evoked potentials (MEPs) produced by transcranial a seizure-free outcome after surgery was the presence of magnetic stimulation to study the physiology of Rolandic a discrete unilateral EEG spike focus. That study also spikes. Among patients with benign Rolandic epilepsy, found that the best outcome after surgery occurred when somatosensory-evoked potentials elicited by stimulation the ictal EEG origin was congruent with a magnetic of the ®ngers were of extremely high amplitude and resonance imaging (MRI) focal lesion. Of patients with appeared as an evoked Rolandic spike. Approximately this arrangement 43% became seizure-free and only 11% 60% of electrical stimuli of the thumb evoked Rolandic were less than 75% improved; when EEG and MRI data spikes among such patients. Using electrical stimulation were less congruent, 48% were still seizure-free but 34% of the thumb as the conditioning stimulus and MEPs by received less than a 75% improvement. Similarly, in transcranial magnetic stimulation as the conditioned paediatric patients, Paolicchi et al. [6] found that stimulus, in control subjects prior electrical stimulation complete resection of not only the epileptogenic lesion, of the thumb inhibited the MEP at interstimulus but also the electrographically abnormal region was the (conditioning±conditioned) intervals of 50±80 ms, main determinant of outcome after focal resections. whereas the MEP in this interval was facilitated among patients with benign Rolandic epilepsy. The facilitating Libenson et al. [7] found approximately the same clinical interstimulus intervals corresponded to the late part of correlations of occipital epileptiform discharges in the ascending phase and the peak of the evoked spike, children as did Smith and Kellaway [8] in 1964! The but not to earlier and later components. The authors majority of the group studied by Libenson et al. and that reasoned that afferent inputs from digital cutaneous by Smith and Kellaway had seizures, but a substantial territories drive a hypersynchronous activation of a large portion remained cryptogenic, with the others suffering population of sensory neurones that are the anatomical various insults at early age; 40% of each group had source of the interictal spikes. A depolarizing event in generalized tonic-clonic attacks, whereas simple and the sensory area could produce a hypersynchronous complex partial attacks of varying descriptions consti- neuronal discharge that is transmitted to motor cortex by tuted the remainder of the seizures in each group. A cortical±cortical connections. This sequence would ex- substantial proportion of each group had cognitive and/or plain the enhancement of motor cortex output corre- behavioural abnormalities, and a signi®cant minority had sponding to the ascending phase and peak of the spike and not to the earlier, presumably sensory components. This study illustrates that valuable neurophysiological Table 1. Localizing value of ictal scalp electroencophalography by insights can be obtained by carefully designed and lobe of seizure origin executed clinical investigation. Epilepsy lobe Seizures [n (%)] Patients [n (%)] Electroencephalography in temporal Frontal 26 (23%) 11 (42%) Lateral temporal 96 (52%) 29 (74%) lobe epilepsy Parietal 4 (10%) 1 (13%) The value of the interictal spike in localizing temporal Occipital 41 (70%) 12 (92%) epileptogenesis has been increasingly recognized in Adapted from Lee et al.[4.]. recent years, as reviewed by So [11 .]. Blume et al. [12] Current trends in electroencephalography Blume 195 found that most or all seizures arose ipsilaterally to the Intracranial electroencephalography majority of temporal spikes in 99 out of 104 patients The aforementioned correlations represent some of the (95%); of those with greater than 3:1 side:side spike factors that have reduced the proportion of patients ratio, this proportion rose to 79 out of 80 patients. Focal undergoing invasive electroencephalography. For ex- unilateral temporal EEG delta activity also was ample, only 21 out of 184 consecutive temporal lobe correlated with seizure origin in over 90% of cases. patients in the patients studied by Cendes et al. [14] Pataraia et al. [13] found that, when temporal lobe underwent such recording. spikes predominated on the same side of hippocampal atrophy, approximately 90% of recorded seizures arose Therefore, such investigation can be avoided when ipsilaterally. The review of So [11 .] indicated that interictal and ictal scalp EEG, and interictal discharges may also predict ef®cacy of anterior neuropsychological testing favour one area of temporal temporal lobectomy, in that congruency between epileptogenesis. As reviewed by Diehl and Luders [16], temporal spike location and ictal onset was associated less congruence likely will require invasive recordings, with excellent postoperative seizure control. A con- i.e. when: (a) EEG monitoring and MRI ®ndings are gruent MRI abnormality enhanced further the prob- discrepant, (b) multifocal seizure onsets are suspected, ability of an excellent surgical outcome to almost 95%. and (c) the site of seizure onset cannot be localized by However, temporal spikes may not appear in approxi- scalp EEG. In nontemporal epilepsy, invasive recordings mately 10% of patients with intractable temporal are frequently required when epileptogenesis from a epilepsy. region distant to surface EEG electrodes is suspected, such as an inferior or mesial frontal or occipital surface Further support of the value of interictal EEG is derived [17]. from a study by Cendes et al. [14]. Those investigators studied lateralization of temporal lobe interictal spikes, Rapid advances in noninvasive technology have resulted hippocampal atrophy on MRI and seizure lateralization in a reduced need for invasive monitoring in children also in 184 consecutive patients with [18 .]. Jayakar [18 .] doubts that invasive monitoring is lesions that were compatible with mesial temporal necessary when all noninvasive data re¯ect an epilepto- sclerosis. Only 3% of patients had discordant interictal genic region within the area of the planned resection or and ictal EEG lateralizations. Lateralization of hippo- when intra-operative (EcoG) reveals campal atrophy agreed closely with that of interictal and almost continual focal discharges. Moreover, that inves- ictal EEG (Cohen k 0.90), and all patients with unilateral tigator cautioned that limitations of sampling and inter- hippocampal atrophy had concordant EEG lateralization. pretation still exist for invasive monitoring. He indicated However, six out of 33 patients with bilateral asymme- the greatest bene®t of invasive monitoring may be trical hippocampal atrophy had discordant EEG later- achieved among children who are neurodevelopmentally alization. intact and who have a restricted epileptogenic zone, as de®ned by scalp EEG and imaging. Patients with clinical Moser et al. [15 .] compared the usefulness of presurgical evidence of a diffuse or multifocal and EEG, MRI and neuropsychological data in the later- diffuse epileptogenic patterns on scalp EEG and multi- alization of temporal lobe epilepsy in patients who focal neuroimaging lesions are unlikely to bene®t. subsequently underwent effective temporal lobectomy. EEG was evaluated as seizure lateralization, MRI as The group of Jayakar found invasive monitoring helpful left±right differences in hippocampal volume, and when studies suggest epileptogenesis in the posterior neuropsychological data consisted of mean scores of ®ve temporal occipital region encroaching upon language cognitive tests. EEG was slightly more reliable for cortex, and when distinction between anterior temporal lateralization than was MRI (89% versus 86%) and both and orbital frontal epilepsy is not clear. Finally, Jayakar were superior to neuropsychological data (66%) using [18 .] and others have noted that the epilepsy associated these measures. with an epileptogenic lesion is not homogeneous in its vicinity. The epileptogenic lesion can be larger or Taken together, data from these studies suggest that smaller than the lesion or may involve one of its interictal temporal lobe spikes and recorded seizures boundaries preferentially [19]. each can lateralize the temporal seizure focus. If they are consistently lateralized to the temporal lobe that Electroencephalography and functional harbours unilateral hippocampal atrophy, or epilepto- neuroimaging genic lesion, then in-patient monitoring could be In addition to new studies using traditional methods for obviated. Patients with scarce interictal spikes or localizing epileptogenesis, functional imaging holds bilaterally appearing indices of epileptogenesis may some promise for delineating epileptogenic areas. These require in-patient investigation. may obviate the need for invasive recording in some 196 Seizure disorders patients, whereas in others it may direct ing the area of seizure onset, as de®ned by invasive ictal placement to the most likely regions, saving the patient EEG. Review of ¯umazenil PET by these authors found unnecessary implantation. correct epileptogenesis localization in 57±100% of studies carried out from 1993 to 1998. Electroencephalography-linked functional magnetic resonance imaging (fMRI) Summary of functional imaging modalities Schomer et al. [20] described their ®ndings from Each of these methods help to localize epileptogenesis functional MRI acquisitions triggered by EEG spikes, in circumstances that are occult to conventional means. having developed instrumentation that allows EEG In common with surface and invasive EEG, each has a within the MRI unit. The relationship between focal sampling problem. Although these methods partially increases in neuronal ®ring rates, as re¯ected in the EEG overcome the spatial sampling limitation of surface and spikes and provoked focal increases in blood ¯ow, invasive EEG, they fail to detect the chronology of the enabled those investigators to develop functional MRI- epileptic process in terms of origin and early propagation. linked images that depict the origin of such activity. Practical and economic considerations limit the duration of such monitoring. One or more of these relatively Magnetoencephalography expensive methods may actually save money if invasive As magnetic signals are far less distorted by intervening monitoring is avoided in some patients. However, only tissues than are electrical signals, magnetoencephalogra- neurophysiological recordings will provide the critical phy (MEG) provides a better spatial resolution of signals temporal resolution for the investigation of seizure than does conventional EEG [21]. MEG dipole model- initiation and propagation in human epilepsy. ling can noninvasively attribute spike activity to subcompartments in a temporal lobe, such as the Electroencephalography in mediobasal temporal lobe and the temporal tip cortex. `The EEG is underused in coma' [24]. Although clinical Moreover, MEG spike dipoles can be localized adjacent examination can assess -stem function in comatose to lesions that are visible on MRI scans [21]. Unfortu- patients, cortical activity is largely closed to clinical nately, MEG requires the patient to remain in a single assessment. A wide spectrum of EEG ®ndings may be position throughout the recording, limiting its duration; seen in patients with a similar impairment of brain-stem ictal recordings are not yet practical. EEG can detect all function. As the variety, complexity and reactivity of MEG-recorded spikes, but some EEG spikes are missed EEG rhythms correlate inversely with the severity of by MEG [21]. The practical value of this approach for cortical dysfunction, certain EEG patterns contribute to detecting epileptogenesis remains uncertain. prognosis for survival. In particular, these involve EEG `suppression' or lack of function, as well as lack of Single-photon emission computed reactivity [25]. The presence of one of several character- tomography (SPECT) istic EEG coma patterns may enable one to determine As reviewed by So et al. [22 .], interictal single-photon the general category of a disease process (i.e. metabolic, emission computed tomography still has low sensitivity drug-induced, or structural). Continuous EEG monitor- (28±66% among various studies) and speci®city in the ing is principally indicated in patients with an unstable detection of an epileptogenic area. However, longer half- or potentially treatable condition such as status epilepti- lives of newer radiotracers have augmented the sensi- cus, recurrent seizures, and variable tivity of detecting ictal hyperperfusion to 90% for a or cerebral perfusion [24]. Because the treatment of temporal focus and to 81% for an extratemporal focus, refractory generalized convulsive seizures may involve and speci®cities are 77 and 93%, respectively [22 .]. The pharmacological paralysis of motor function, electroen- group of So et al. [22 .] superimposed the subtraction of cephalography is necessary to assess progress. Noncon- interictal (hypoperfusion) from ictal (hyperperfusion) vulsive may develop de novo or may scans on MRI, and improved the sensitivity for detecting evolve from convulsive status epilepticus. By performing an epileptogenic focus to 88%. Because standard MRIs routine EEGs on comatose patients, Towne et al. [26] disclosed no lesion in half of these patients, such found nonconvulsive status epilepticus in 8% of 236 ®ndings hold considerable promise for nonlesional patients with no overt clinical seizure activity. Age and intractable epilepsy. aetiology did not distinguish patients with and without nonconvulsive status epilepticus. These authors found Positron emission tomography (PET) generalized, hemispheric, or bilateral independent per- In addition to the traditional ¯uorodeoxyglucose, posi- sistent epileptiform activity. tron emission tomography (PET) may also detect altered receptor function, speci®cally ¯umazenil Conclusion imaging. JuhaÂsz et al. [23] found ¯umazenil PET to be Mysteries and complexities of more sensitive than ¯uorodeoxyglucose PET in detect- function and dysfunction are best unravelled by studies Current trends in electroencephalography Blume 197 and investigations combining methodologies. Therefore, 11 So EL. Integration of EEG, MRI, and SPECT in localizing the seizure focus for . epilepsy surgery. Epilepsia 2000; 41(suppl 3):S48±S54. it is gratifying to see in this age of prominent This is a useful review comparing three methods of localizing epileptogenesis. neuroimaging, that clinical electrophysiology retains its 12 Blume WT, Borghesi JL, Lemieux JF. Interictal indices of temporal seizure essential role. origin. Ann Neurol 1993; 34:703±709. 13 Pataraia E, Lindinger G, Deecke L, Baumgartner C. Whole-head MEG in presurgical epilepsy evaluation. In: 3rd European Congress of Epileptology, May 24±28, 1998, Warsaw, Poland. Majkowski J, Owczarek K, Zwolinski P References and recommended reading (editors). Bologna: Monduzzi Editore; 1998. pp. 135±139. Papers of particular interest, published within the annual period of review, have 14 Cendes F, Li LM, Watson C, et al. Is ictal recording mandatory in temporal been highlighted as: lobe epilepsy? Not when the interictal electroencephalogram and hippocam- . of special interest pal atrophy coincide. Arch Neurol 2000; 57:497±500. .. of outstanding interest 15 Moser DJ, Bauer RM, Gilmore RL, et al. Electroencephalographic, volumetric, . and neuropsychological indicators of seizure focus lateralization in temporal 1 Binnie CD, Stefan H. Modern electroencephalography: its role in epilepsy lobe epilepsy. Arch Neurol 2000; 57:707±712. . management. Clin Neurophysiol 1999; 110:1671±1697. Their data also support the value of surface electroencephalography in lateralizing This review provides sensitivity and specificity data which could be used in positive temporal lobe epilepsy, supporting the diminishing need for invasive electro- predictive value studies. encephalography. 16 Diehl B, Luders HO. Temporal lobe epilepsy: when are invasive recordings 2 Ajmone Marsan C, Zivin LS. Factors related to the occurrence of typical needed? Epilepsia 2000; 41(suppl 3):S61±S74. paroxysmal abnormalities in the EEG records of epileptic patients. Epilepsia 1970; 11:361±381. 17 Zumsteg D, Wieser HG. Presurgical evaluation: current role of invasive EEG. Epilepsia 2000; 41(suppl 3):S55±S60. 3 Gilbert DL, Buncher CR. An EEG should not be obtained routinely after first unprovoked seizure in childhood. Neurology 2000; 54:635±641. 18 Jayakar P. Invasive EEG monitoring in children: when, where, and what? J Clin . Neurophysiol 1999; 16:408±418. 4 Lee SK, Kim JY, Hong KS, et al. The clinical usefulness of ictal surface EEG in to principles elaborated by Jayakar will guide clinicians in selecting . neocortical epilepsy. Epilepsia 2000; 41:1450±1455. children most appropriate for invasive monitoring. This study indicates that surface EEG can identify epileptogenesis, particularly for occipital and temporal lobe epilepsy. 19 Pathak P, Blume WT. The asymmetrical epileptogenicity of brain lesions [abstract]. Can J Neurol Sci 1997; 24(suppl 1):S15±S16. 5 Holmes MD, Kutsy RL, Ojemann GA, et al. Interictal, unifocal spikes in 20 Schomer DL, Bonmassar G, Lazeyras F, et al. EEG-linked functional . refractory extratemporal epilepsy predict ictal origin and postsurgical outcome. magnetic resonance imaging in epilepsy and cognitive neurophysiology. J Clin Neurophysiol 2000; 111:1802±1808. Clin Neurophysiol 2000; 17:43±58. This study provides additional supportive data for the value of interictal EEG, the most common situation presented to the clinician. 21 Baumgartner C, Pataraia E, Lindinger G, Deecke L. Neuromagnetic recordings in temporal lobe epilepsy. J Clin Neurophysiol 2000; 17:177±189. 6 Paolicchi JM, Jayakar P, Dean P, et al. Predictors of outcome in pediatric epilepsy surgery. Neurology 2000; 54:642±647. 22 So EL, O'Brien TJ, Brinkmann BH, Mullan BP. The EEG evaluation of single . photon emission computed tomography abnormalities in epilepsy. J Clin 7 Libenson MH, Caravale B, Prasad AN. Clinical correlations of occipital Neurophysiol 2000; 17:10±28. epileptiform discharges in children. Neurology 1999; 53:265±269. This communication describes Subtraction Ictal Single Photon Emission Computed Tomography Coregistered to Magnetic Resonance Imaging (SIS- 8 Smith JMB, Kellaway P. The natural history and clinical correlates of occipital COM), a potentially useful technique in localizing epileptogenesis when MRI is foci in children. In: Neurological and electroencephalographic correlative normal. studies in infancy. Kellaway P, Petersen I (editors). New York: Grune & Stratton; 1964. pp. 230±249. 23 Juha sz C, Chugani DC, Muzik O, et al. Relationship between EEG and positron emission tomography abnormalities in clinical epilepsy. J Clin 9 Kellaway P. The electroencephalographic features of benign centrotemporal Neurophysiol 2000; 17:29±42. (rolandic) epilepsy of childhood. Epilepsia 2000; 41:1053±1056. 24 Young GB. The EEG in coma. J Clin Neurophysiol 2000; 17:473±485. 10 Manganotti P, Zanette G. Contribution of motor cortex in generation of evoked 25 Young GB, Kreeft JH, McLachlan RS, DeMelo J. EEG and clinical .. spikes in patients with benign rolandic epilepsy. Clin Neurophysiol 2000; associations with mortality in comatose patients in a general intensive care 111:964±974. unit. J Clin Neurophysiol 1999; 16:354±360. This is a particularly carefully done, multi-modal neurophysiological study exploring the pathophysiology of the most common form of benign focal of 26 Towne AR, Waterhouse EJ, Boggs JG, et al. Prevalence of nonconvulsive childhood. Its value lies in both the careful methodology and in the results. status epilepticus in comatose patients. Neurology 2000; 54:340±345.