EEG Source Localization of Interictal and Ictal Spikes Acquired During Video-EEG Monitoring

Transcription

1 Ahmed A. Gaber et al. EEG Source Localization of Interictal and Ictal Spikes Acquired During Video-EEG Monitoring Ahmed A. Gaber, Youssry A. Abo El Naga, Ahmed M. Hazzo, M. Ossama Abdulghani, Khaled O. Abdulghani Department of Neurology, Ain Shams University ABSTRACT Objective: We investigated the feasibility of electroencephalography (EEG) dipole source localization of interictal and ictal epileptiform discharges from data acquired during routine clinical video-eeg monitoring (VEM) that utilizes a 25- channel 'routine montage' with an additional inferior electrodes 'surgical montage'. Methods: Twenty five consecutive patients who had VEM for the presurgical evaluation of medically refractory partial epilepsy were screened. EEG data was acquired with 25 electrodes including an inferior row (surgical montage). For comparison, the 10 additional electrodes were excluded from analysis (routine montage). Using ASA 2 software, a computed dipole source localization of averaged spikes was performed utilizing a magnetic resonance imaging-based finite element model. Dipole localization was compared with that of the comprehensive epilepsy program evaluation. Results: Dipole source localization using ASA was highly complementary to the four landmarks of the epileptogenic zone, the ictal symptomatic zone, irritative zone, pace maker zone and epileptogenic lesion. The concordance between these zones was 32% but with ASA it reached 96% with highly significant difference. The lateralizing ability of data without ASA was not statistically significantly different after the use of ASA (P=0.247). As regards the localization ability before ASA only 4 cases could be localized and 14 cased had regional localization, after ASA all cases could be localized with highly significant difference (P= ). Conclusions: EEG source localization has a significant diagnostic yield in localization of epileptogenic zone in preoperative settings, however this yield needs to be validated by invasive EEG. (Egypt J. Neurol. Psychiat. Neurosurg., 2006, 43(1): ) INTRODUCTION Drug resistant focal epilepsy is responsible for high social and economic costs. 1 Approximately 60% of all patients with epilepsy suffer from focal epilepsy, and in one third of these patient seizures are not adequately controlled with antiepileptic drugs. 2 Surgery is currently accepted as an effective and safe therapeutic approach in drug resistant epilepsy, particularly in lobe epilepsy, where patients became seizure free in 70-90% of cases. 3 However, surgery in epilepsy remain an underused last resort treatment because only a small proportion of patients affected by surgically remediable epilepsies undergo this intervention. In a recent editorial, Engel 4 reported fear of morbidity and confidence of new antiepileptic drugs or in vagal nerve simulation as factors that discourage patients and their physicians from surgery, which lack sufficient data from randomized controlled trials. However, a recent randomized controlled trial of lobe epilepsy surgery found it to be superior to prolonged medical treatment in terms of efficacy and safety. 5 Moreover, although today there is agreement that lobe epilepsy can be diagnosed in most patients without invasive tests 6-8, there is no consensus between different epilepsy surgery groups regarding the optimal use of non invasive procedures. The aim of an ideal non invasive presurgical protocol should be to; (1) identify the epileptogenic zone and consequently to identify potential candidates for invasive recording, (2) minimize the cost of human and technological resources, and (3) reduce the time of diagnostic evaluation, so that surgery can be offered to more patients

2 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 We implemented a non invasive presurgical evaluation protocol for intractable epilepsy cases, based on "antamo-electro-clinical correlations", inspired by methodological principles first proposed by Bancaus and Talairach, 10 and later developed by their successors In many cases, complicated generator configurations underlie EEG recordings, involving multiple focal or extended sources. Therefore, it is advantageous to apply several methods to the same data set in order to obtain converging evidence or to draw conclusions from different results yielded by different methods. 14 The power of EEG is mostly in its properties in space. The resolution itself is one or two orders of magnitude better than in the imaging methods. EEG is also basically considered as a signal, not an image, and most of the interpretation is based on the waveform analysis. Here comes the need to improve its spatial resolution by combining it with anatomical data. The first trial to combine electrophysiological data and anatomical data was the use of the scalp EEG maps. However there were limitations in these techniques, the most important of which is the scalp EEG does not represent actually the intracranial generators of the EEG. The advent of the MRI was utilized to provide the anatomical landscape to compute the intracranial generators i.e. to compute a realistic intracranial maps. 22 There are several levels for the utilization of MRI information in EEG studies. By increasing the complexity level more specific information is obtained. The complexity involves processing of both the EEG and MRI data. Various processing methods are described in figure (1) as well as the resulting combinations. 22 Fig. (1): The use of MRI in EEG studies

3 Ahmed A. Gaber et al. In the process of localization of source activity, because the electrophysiology of the brain is very complicated, The EEG is roughly divided into event-related potentials and spontaneous activity. Both are distributed in a wide range of the brain, but especially some periods of event-related potentials are focused in certain locations. Also certain abnormal activity, such as epileptic spikes, may originate focally in the brain. These cases fulfill the conditions of the traditional source analysis. It is based on the assumption that the EEG can be described by one electrical point dipole source at a certain moment or period of time. More advanced methods imply multiple dipoles and distributed sources, which are in most cases more realistic models for the electrical brain activity. 23 Source analysis implies the solution of the inverse problem, which is based on several solutions of the forward problem. Not only the source has to be modeled, but also the head must be replaced by a volume conductor model. 24 This model is in all practical solutions linear, resistive, piece-vice homogeneous and isotropic. In principle it would be possible to deal with even more realistic models, which have anisotropic and complex impedances. Further, the minimum size of each compartment in the model could be the dimension of the voxels obtained from the MRI scan. However, all these features demand lot of computing power and are therefore not included in present head models. The forward models of the head can be divided into spherical 25,26 and anatomical realizations. Spherical models work fast and are thus optimal for patient studies consisting of large materials. Anatomical models are time-consuming, but give the individual geometry of the subject and suit for special cases. Although MRI is not needed for the computation in the former case, the MRI scans are often used also in spherical modeling. The sphere can be positioned either in the center of the head or asymmetrically so as to produce an optimal source localization. 32 Independently on the modeling method, the MRI images are often used for displaying the anatomical locations of the sources. Truly accurate dipole source localization relies on having; (1) reliable, artifact-free EEG data to start with, (2) landmark and electrode position data obtained with a three-dimensional digitizer, (3) MRI data from the same subject (with a means to measure the same landmarks), (4) the capability to apply different dipole and volume conductor models, and (5) the ability to coregister the functional and anatomic data to display the final source solutions. Most important is to have a thorough understanding of the fundamental principles of the localization programs and neurophysiologic processes. Localization programs are simply tools. Used incorrectly, they can yield meaningless results. Used correctly, they can provide a window into neurophysiologic functioning that is not available through any other means. 15 So we investigated the feasibility of electroencephalography (EEG) dipole source localization of interictal and ictal epileptiform discharges from data acquired during video-eeg monitoring (VEM) that utilizes a 25-channel 'routine montage' with an additional inferior electrodes 'surgical montage. 15 The aim of this work was to assess the concordance rate between the scalp EEG ictal and interictal and the anatomical data provided by the MRI as well as MRS with those provided by source analysis techniques, and to assess the diagnostic yield of dipole source localization as a tool for non invasive localization of epileptic focus. PATIENTS AND METHODS Between the period of May 2002 and January 2005, twenty one cases of medically intractable epilepsy were enrolled consecutively in the study. Inclusion criteria included medically intractable partial epilepsy (clinical and/or electrophysiological) with a follow up of at least one year excluding generalized epilepsy and generalized epileptic syndromes. This exclusion criterion was made to assess the validity of focus localization by advanced source analysis. There were no age limit or seizure semiology restrictive criteria. Clinical Evaluation: A detailed medical history was obtained from all patients as a first step. Particular attention was paid to recognizing the symptoms/signs at seizure 303

4 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 onset and auras, because it is largely accepted in the literature that the initial seizure semiology usually provides valuable information about the seizure onset zone A full clinical general and neurological examination was carried out in all patients. EEG monitoring: EEG was recorded using an Epineuro Mizar 40 channel system with sampling rate of 256 Hz. All cases underwent video EEG monitoring for at least 8 hours. EEG data was acquired with 25 electrodes using the system and including in addition an inferior row (surgical montage). Ictal recording was not a necessity in inclusion in the study except in those cases that were decided for surgical intervention. The whole interictal EEG in the awake state, including hyperventilation and photic stimulation, and all sleep recordings were evaluated to assess the presence of background abnormalities, interictal slow activity, and epileptiform activity; focal (over one to three channels), regional (over three or more channels), hemispheric (over all channels of one side), and diffuse (over all channels of both sides). 19 All patients admitted to video-eeg monitoring had at least one seizure recorded. Ictal EEG findings: We categorized ictal events as; (1) seizures with subjective phenomena only (auras), and (2) seizures with symptoms and/or signs, with or without loss of contact (simple partial or complex partial seizures). Each ictal event was correlated with the corresponding EEG changes, when present. According to the site, we distinguished between focal, regional, hemispheric, or diffuse discharges. According to the EEG seizure pattern appearance and course, ictal EEG changes were classified into ictal onset pattern (first sudden change of frequency with attenuation or appearance of a new rhythm), ictal core pattern, late patterns, and post ictal pattern. 9 Neuroradiological evaluation: The patients underwent brain 1.5 Tesla magnetic resonance imaging (MRI) examinations. FLAIR technique was an essential criterion in our study for visual assessment of the hippocampus. Four cases had a magnetic resonance spectroscopy done for more accurate epileptogenic zone assessment. The MRI scans were reviewed by neuro-radiologists, masked to the clinical and outcome data. At review, MRI data were classified as follows: mesial sclerosis (MTS), lesional (as low grade tumors) or non lesional. The presence of MTS was evaluated qualitatively by visual inspection of MRI. Atrophy (on T1 weighted sequences) and increased mesial signal intensity (on T2 weighted and fluid attenuated inversion recovery sequences) were considered markers of MTS. Volumetry was not performed. 20,21 Dipole Localization with Advanced Source Analysis Technique: The advanced source analysis was done using the ANT software system ASA (Advanced Source Analysis) (version 2). In this study, the recording of an epileptic spike in analyzed first by the multiple signal classification (MUSIC) method, which allows the detection of multiple focal generators with non-coherent time courses, and second by a constrained fixed dipole model. An anatomical MRI of some patients was not available, neither were any exact positions of the electrodes. So a standardized realistic head model was used to generate solutions in the Talairach co-ordinate space. The electrode locations on the scalp were taken according to the standard. From the ongoing EEG, a piece of 250 ms centered on each spike was selected. In order to come to a first estimation of active regions in the brain, the MUSIC method is applied to the selected time window. Then, a constrained dipole fit scheme was carried out. We used a single dipole method and we compared dipole localizations of both averaged and individual spikes. The ASA spatial ability for lateralization and localization is different from that in EEG as it utilizes both the electrophysiological and anatomical data for localization. The data of ASA is presented as dipole which is the unit of source generator in an anatomical area with a three dimensional construct. So each dipole has an X,Y,Z related to the selected landmarks on the MRI. Thus the term focal activity here means a single or multiple dipoles in a localized 304

5 Ahmed A. Gaber et al. anatomical area even in the diencephalon. More than one focus is considered focal not regional as it is well localized to a small anatomical structure. Regional is considered if multiple dipoles are presented in contagious areas for example mesial and neocortical. A generalized activity is considered if multiple dipoles are located on either sides of the midline. Classification of results: Results were classified according to the delineation of the epileptogenic zone proposed by Luder et al. 33 : 1. Ictal symptomatic zone: The epileptogenic zone defined by seizure semiology. 2. Irritative zone: The epileptogenic zone defined by the interictal EEG. 3. Pace maker zone: The epileptogenic zone defined by the ictal EEG. 4. Epileptogenic Lesion: The anatomical abnormality evident by the MRI. A fifth category was added which is the definition of the epileptogenic zone using source analysis and dipole localization. The concordance of these results was assessed. They were categorized according to the concordance between the four categories of Luder regardless the results of source analysis to three groups: 1. Completely concordant: all the four categories of the epileptogenic zone are present and concordant. 2. Partially concordant: at least three of the four main categories are concordant. 3. Discordant. The addition of the source analysis results to these data was evaluated for its effectiveness to give either more concordant or discordant results. Scheme for localization of Partial epilepsies: Patients were classified into either lobe or extra lobe epilepsy. EEG ictal onset in extra regions were categorized as frontal, occipital or parietal depending on ictal clinical findings. Ictal onset semiology characterized by focal motor, sensory, visual phenomena, or complex motor manifestations (forced head turning or hyper motor behavior) suggests a non- lobe seizure onset. In patients with lobe epilepsy (TLE), we used a classification of EEG ictal core patterns according to well known studies concerning ictal scalp intracranial recordings, which showed a strict correlation between ictal scalp morphology of discharge and different lobe structures involved at seizure onset, as follows: Type 1: antero 5 9 Hz discharge, associated with a highly probable onset in the mesial structures. Type 2: 2 4 Hz discharge, associated with a highly probable onset in the lateral structures. In implementing the TLE diagnostic grid, despite the fact that no ictal behavior is specific for lobe seizures, we considered some initial symptoms/signs as highly suggestive of seizures and classified them into a mesial or lateral cluster 9 (Table 1). Table 1. Differentiating clinical, anatomical and electrophysiological criteria for identification of lobe epilepsy subsyndromes. 9 Mesial cluster Lateral cluster Anatomical criteria Ictal EEG criteria Ictal clinical criteria Mesial sclerosis or hippocampal atrophy and/or medial structural lesion 5-9 Hz discharge, well localized to regions. Rising epigastric aura, isolated or associated with other vegetative symptoms. Early oroalimentary automatism Dystonia of contralateral upper limb. Focal atrophy of lateral neocortex and/or lateral structural lesion. 2-5 Hz ictal discharge well localized to regions Auditory or vestibular symptoms. Staring without oroalimentary automatism. 305

6

7 Ahmed A. Gaber et al. Patients with TLE were subdivided into the following three groups according to Quarato et al. 9 : * Mesial TLE: when the anatomical criterion for mesial cluster, the EEG criterion for mesial cluster, and at least two of the clinical criteria for mesial cluster were met. * Lateral TLE: when the anatomical criterion for lateral cluster, the EEG criterion for lateral cluster, and at least two of the clinical criteria for lateral cluster were met. * Mesiolateral TLE: when at least one criterion for mesial cluster and at least one criterion for lateral cluster were met, or when the anatomical and the EEG criteria were concordant with one clinical criterion alone. If brain imaging was negative, only the presence of EEG and clinical criteria were taken into account for the classification. Statistical analysis Descriptive statistics were used to summarize the data for all variables. The Chi-square test with Yates correction and Fischer exact test were used to test for differences in diagnostic yield of different assessment modalities. RESULTS Twenty Five cases were enrolled in this study. Fifteen were males and ten were females with a mean age of years. All cases were medically intractable epilepsy fulfilling the criteria of medical intractability 33 (although it was not a prerequisite for this study) that underwent extensive non invasive neurophysiological evaluation as a part of preoperative evaluation. The mean seizure duration in those patients was years. Clinical evaluation Medical intractability was assured in all cases. An exhaustive search for the ictal symptomatic zone was a main target. According to the ictal symptomatic zone (i.e. seizure semiology) the cases were preliminary classified as and extra epilepsy. Out of the 25 cases, 17 cases were (68%) (11 mesial, 1 lateral, 5 mesiolateral), 5 extra (20%) (4 frontal, 1 hypothalamic), 3 mixed and extra lobe epilepsy (12%) (Table 2). Neurophysiological evaluation Interictal and ictal EEG results were rated by two raters blinded to each other. The EEG recorded was assessed mainly for its spatial ability to localize the epileptogenic zone. This ability was assessed into two axes, first its lateralizing value (the ability to lateralize the hemisphere of epileptogenic origin), and second the localizing value (the ability to localize the epileptogenic zone in the hemisphere). The localizing ability was rated by whether the EEG findings were focal, regional, hemispheric, generalized, diffuse or uneventful. If the recording showed two foci whether unilateral or bilateral it is considered regional. Interictal EEG findings The interictal EEG lateralized 15 cases (60%), while 4 cases (16%) were bilateral and 6 cases (24%) were non lateralized. Those non lateralized cases were uneventful (3 cases) and with generalized interictal EEG findings (3 cases) (Tables 3, 4). Out of the 25 cases only 5 cases were focal (20%), 14 regional (56%), 3 generalized (12%) and 3 uneventful (12%). One of those regional cases had two focal generators (mesial and neocortical ) (Tables 3, 5). Ictal EEG findings All cases had ictal EEG findings at least for one seizure with an average of 1.6 seizures. The ictal EEG lateralized 19 cases (68%), 3 cases (12%) were bilateral and 3 (12%) could not be lateralized (all of them had generalized activity) (Table 3, 4). There was no statistically significant difference between the interictal and ictal EEG as its lateralizing value (P value = 0.7) The ictal EEG was able to localize 6 cases (24%) as focal, 16 cases (64%) as regional and 3 cases (12%) as generalized. In four of those regional cases a focal onset could be detected but 307

8 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 with a different generator. Three of them had mesial as well as neocortical generators. One of them had one mesial and another diencephalic generator (Tables 3, 5). There was no statistically significant difference between the interictal and ictal EEG as its Localizing value (P value = 0.67). ASA findings: The ASA when applied to the recorded EEG (ictal and interictal) lateralized 23 cases (92%), and 2 cases (8%) were not lateralized, both had a midline generator. All 25 cases could be localized. Single generator was found in 21 cases, 2 of them were midline (considered previously by the ictal EEG as generalized). Two generators were found in four cases who were considered by the ictal EEG as regional. The localizing and lateralizing ability of ASA was not affected whether applied to interictal or ictal EEG. However the goodness of fits is affected by the number of recorded spikes. So the inclusion of ictal recording makes us more sure of the results. By comparing the results of ASA (Tables 4, 5; Fig. 2, 3) using Chi -square test with Yates correction there was no statistically significant difference between the ASA and the ictal EEG results as regards the lateralizing ability (P=0.247). However there is statistically significant difference between the lateralizing ability of ASA and interictal EEG (P=0.02). The localizing ability of ASA was highly significant superior to both interictal (P= ) and ictal EEG (P= ). Neuro-imaging findings All cases had MRI brain while only 4 cases had magnetic resonance spectroscopy (MRS). The MRI findings were positive in 22 cases (88%) and negative in only 3 cases (12%). Mesial sclerosis (MTS) was found in 15 cases (60%), mesial astrocytoma in 2 cases (8%), hypothalamic hamartoma in 2 cases (8%), AVM in one case (4%), encephalomalacia in one case (4%) and non specific hyperintensity in one case (4%). One of the cases with mesial sclerosis had hemiatrophy as well (Table 3). Four cases underwent Magnetic resonance spectroscopy (MRS). Two cases of them had normal MRI and MTS was found after MRS. The other two cases had bilateral MTS and MRS confirmed that but revealed a predominant side (Table 3). Integration of the clinico-electro-antaomical data; The epileptogenic zone The data of the above assessment were then integrated (Table 6). According to these data we had 17 mesial lobe epilepsy, 1 mesiolateral and 1 neocortical lobe epilepsy, 3 mixed and extra lobe epilepsy, and 3 extra (1 frontal lobe epilepsy and 2 of hypothalamic origin). The concordance of the data was assessed before and after the use of source analysis techniques (Tables 7, 8; Fig. 4). The use of source analysis added a new information or confirmed previous information and pushed the data toward concordance. Before source analysis concordant data was achieved only in 8 cases (32%), while after source analysis it became 24 (96%). There was a highly significant statistical difference between the concordance of data before and after ASA (P=0.0001). The lateralizing ability of the data was assessed before and after ASA (Table 9). Six cases were non lateralized before ASA, but after ASA only 2 cases were non lateralized as they had mid line generator, and were proved anatomically and clinically to be hypothalamic hamartoma. However this was not statistically significant (P=0.247). The localizing ability of data without and with ASA was assessed (Table 10, Fig. 5). Before the use of ASA only 4 cases could be localized while in 14 cases regional localization to an area could be done, while 7 cases could not be localized, After the use of ASA all cases could be localized with statistically significant difference as regards the localizing ability of the ASA added to other methods (P= using Fisher s exact test). 318

9 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 Table 2. Demographic data and Semiological classification. ID Age Sex D Semiology Classification 1 28 f 22 left motor SPS in UL & face Frontal 2 45 m 20 CPS dialeptic, rising gastric sense Mesial 3 35 f 22 CPS dialeptic, rising gastric sense Mesial 4 27 f 6 Myoclonic astatic + CPS with G (fear rising gastic Mesial + sensation, dialeptic) Generalized 5 23 m 16 CPS,right adversive fits with G Frontal 6 25 m 24 CPS psychoemotional with G Mesial Temporal 7 8 m 8 CPS jelastc psychoemotional Hypothalamic 8 43 m 10 CPS sexual feeling rising gastric sensation Mesial Temporal 9 28 m 13 CPS with abd aura rising gastric sensation Mesial f 13 CPS dialeptic rising gastric sensation Mesial Temporal f 3 CPS frontal with right SPS motor in UL with bimanual automatism Frontal f 6 CPS with abdominal aura rising gastric sense + neocortical frontal SPS in angle of mouth and UL Mesiolateral m 10 CPS psychoemotional, gelastic Mesial Temporal f 6 Right MTLE abdominal aura and rising gastric sensation Mesial Temporal f 26 CPS dialeptic rising gastric sensation, buccoral automatism, with g (MTLE with G) Mesial Temporal m 14 CPS dialeptic rising gastric sensation, fear, tachycardia, bimanual automatisms with G Mesial Temporal m 20 Right motor SPS faciobrachial with G Frontal m 11 CPS bimanual automatism Mesiolateral m 20 CPS with G right manual automatism Mesiolateral 20 3 m 3 CPS visual hallucination fearful Lateral m 20 CPS with G dejavu followed by left clonic contraction in the UL with G Mesiolateral f 20 CPS with G Mesiolateral m 13 SPS fear, rising gastric sensation, ill defined visual hallucination and secondary generalization Mesial f 3 SPS electric sensation (lt>rt) facio brachial with versive movement of the face with ictal speech arrest + fears to Visual Hallucination (ie SPS Fronto passing to complex with G) m 2 CPS fronto with G aggression with right sided todds Fronto D = duration, f = female, m = male, SPS = simple partial seizures, complex partial seizures, UL = upper limb, LL= lower limb, G = generalization 318

10 Ahmed A. Gaber et al. Table 3. Results of non invasive evaluation of intractable epilepsy patients. ID Interictal EEG Ictal EEG ASA MRI brain MRS 1 Focal R frontal Focal R anterior R mesial R MTS & hemiatrophy 2 Uneventful Regional R R mesial R MTS 3 Regional Focal L mesial bi L mesial L MTS 4 Generalized Regional L L mesial + + diencephalic Diencephalic L MTS 5 Regional Regional bi R mesial R mesial bi R>L R>L astrocytoma 6 L occipital + L L ant + L L ant + L L temprooccipital mesial tempro occipital tempro occipital encephalomalacia 7 Generalized Generalized Diencephalic hypothalamic hamartoma 8 R regional R Regional R mesial free R MTS 9 R Focal anterior bi dependant foci R mesial R MTS 10 R regional R mesial R R mesial astrocytoma 11 L regional L neocortical L posterior neocortical L posterior AVM 12 R mesial R regional R mesial & R and R neocortical frontal rolandic R MTS 13 R regional R mesial and posterior parietal R mesial hyperintense right mesial temprooccipital surface 14 Uneventful R mesial R mesial R MTS 15 L focal L mesial L mesial L MTS 16 Uneventful R mesial R mesial and and diencephalic diencephalic generator Bilateral MTS 17 L focal frontal L regional frontal L insular L insular hyperintensity 18 R regional R anterior R mesial Bilateral MTS 19 L focal L regional L mesial Bilateral MTS 20 Generalized generalized diencephalic generator Hypothalamic hamartoma L regional L focal anterior bilateral regional 24 bilateral L regional L mesial L MTS L regional L mesial L MTS generalized R mesial Bilateral MTS bilateral regional R mesial free R>L MTS L>R MTS L regional 25 L regional L mesial free left MTS R= right, L= Left, U= unlateralized, MTS= mesial sclerosis 309

11

12 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 Table 4. The lateralizing abilities of different electrophysiological methods. Method Interictal EEG Ictal EEG ASA Lateralizing Non lateralizing Bilateral Chi square test with Yates correction comparing ASA and ictal EEG P=0.247 Chi square test with Yates correction comparing ASA and inteictal EEG P= ASA Ictal EEG Interictal EEG 0 lateralizing NonLateralizing Bilateral Fig. (2): The lateralizing abilities of different electrophysiological methods. Table 5. The localizing abilities of different electrophysiological methods. Method Interictal EEG Ictal EEG ASA Focal Regional Generalized Uneventful Chi square test with Yates correction comparing ASA and ictal EEG P= Chi square test with Yates correction comparing ASA and interictal EEG P= ASA Ictal EEG Interictal EEG 5 0 Focal Regional Generalized Uneventful Fig. (3): The localizing abilities of different electrophysiological methods. 318

13 Ahmed A. Gaber et al. Table 6. Epileptogenic Zones. ID ISZ Irritative Zone Pacemaker Zone Epileptic lesion ASA 1 R frontal Rolandic R frontal R anterior R MTS, R R mesial hemiatrophy 2 U Mesial Unknown R R MTS R mesial Temporal 3 U Mesial Temporal Bi L mesial Temporal L MTS L mesial 4 Mesial Temporal & Diencephalic Generalized L, Diencephalic L MTS L mesial & diencephalic 5 R mesial U Frontal Bi R>L Bi R>L astrocytoma R mesial 6 L occipital & L L anterior L temprooccipital L anterior U Mesial Temporal mesial & L tempro occipital encephalomalacia & L tempro occipital 7 hypothalamic Hypothalamic Generalized centriencephalic hamartoma centriencephalic 8 U Mesial Temporal R R R MTS R mesial 9 R anterior R Mesial Temporal Bi R MTS R mesial 10 R mesial U Mesial Temporal R R astrocytoma R mesial 11 L neocortical L posterior L posterior L Frontal perisylvian L AVM neocortical 12 R mesial & R mesial R mesiolateral R R neocortical R MTS and R frontal rolandic 13 U Mesial Temporal R R mesial and posterior parieto hyperintense right mesial temprooccipital surface R mesial 14 R mesial Unknown R mesial R MTS R mesial 15 U mesial L L mesial L MTS L mesial 16 R mesial bilateral MTS more R mesial U mesial Unknown and diencephalic R and diencephalic generator 17 left insular L frontal rolandic L frontal L frontal hyperintensity left insular 18 Temporal R R anterior Bilateral MTS R mesial 19 Bilateral MTS more Temporal L L L L mesial 20 Neocortical Generalized Generalized Hypothalamic hamartoma diencephalic generator 21 L mesiolateral L L L MTS L mesial 22 L anterior U Temporal L L MTS L mesial 23 bilateral U mesial generalized Bilateral MTS R mesial 24 bilateral U Fronto bilateral Free R mesial 25 L fronto L L L MTS L mesial ISZ= ictal symptomatic zone, R= right, L= Left, U= unlateralized, MTS= mesial sclerosis 311

14 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 Table 7. The concordance between different assessment methods. ID Concordance without ASA Concordance after ASA Epileptic Zone 1 Discordant Concordant R mesial 2 Partially concordant Concordant R mesial 3 Concordant Concordant L mesial 4 Partially concordant Concordant L mesial and diencephalic 5 Partially concordant Concordant R mesial 6 Partially concordant Concordant L anterior and L temprooccipital 7 Partially concordant Concordant Hypothalmic 8 Concordant Concordant R mesial 9 Concordant Concordant R mesial 10 Partially concordant Concordant R mesial 11 Discordant Concordant L posterior neocortical 12 Partially concordant Concordant R mesial and right neocortical 13 Partially concordant Concordant R mesial 14 Concordant Concordant R mesial 15 Concordant Concordant L mesial 16 Partially concordant Concordant R mesial and diencephalic 17 Partially concordant Concordant L insular 18 Partially concordant Concordant R mesial 19 Partially concordant Concordant L mesial 20 Discordant Concordant Hypothalamic 21 Concordant Concordant L mesial 22 Concordant Concordant L mesial 23 Concordant Concordant R mesial 24 Discordant Partially concordant R mesial 25 Partially concordant Concordant L mesial Table 8. The concordance without and with ASA. Before ASA After ASA Concordant 8 24 Partially concordant 13 1 Discordant 4 0 Fisher s exact test P= Concordant Partial Concordant Discordant 5 0 Without ASA With ASA Fig. (4): The concordance without and with ASA. 318

15 Ahmed A. Gaber et al. Table 9. Lateralizing and Localizing Value of Dipole source analysis. ID Lateralization without Lateralization with ASA Localization without Localization with ASA ASA ASA 1 Lateralized Lateralized Non localized Localized 2 Lateralized Lateralized Regional Localized 3 Lateralized Lateralized Localized Localized 4 Lateralized Lateralized Regional Localized 5 Non lateralized Lateralized Non localized Localized 6 Lateralized Lateralized Regional Localized 7 Non lateralized Non lateralized Non localized Localized 8 Lateralized Lateralized Regional Localized 9 Lateralized Lateralized Localized Localized 10 Lateralized Lateralized Regional Localized 11 Lateralized Lateralized Non localized Localized 12 Lateralized Lateralized Regional Localized 13 Lateralized Lateralized Regional Localized 14 Lateralized Lateralized Localized Localized 15 Lateralized Lateralized Localized Localized 16 Non Lateralized Lateralized Non Localized Localized 17 Lateralized Lateralized Regional Localized 18 Lateralized Lateralized Regional Localized 19 Lateralized Lateralized Regional Localized 20 Non Lateralized Non lateralized Non localized Localized 21 Lateralized Lateralized Regional Localized 22 Lateralized Lateralized Regional Localized 23 Non lateralized Lateralized Regional Localized 24 Non lateralized Lateralized Non localized Localized 25 Lateralized Lateralized Regional Localized Table 10. The difference between localization ability with and without ASA. Localization Without ASA With ASA Localized 4 25 Regional Localization 14 0 Non Localized 7 0 Fisher s exact test P= Localized Regional Non Localized 5 0 Without ASA With ASA Fig. (5): The difference between localization ability with and without ASA. 313

16 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan 2006 DISCUSSION Determining the precise localization of the primary epileptic focus remains a difficult task. Ictal clinical manifestations can be helpful in predicting the location of the epileptic focus but are not always reliable because clinical manifestations can represent seizure spread from non eloquent regions. 37 Scalp EEG remains an important tool in the localization of the epileptic focus, but the spatial resolution of this technique is low. Ictal singlephoton emission computerized tomography scans can also be helpful; however, spatial resolution is also low and considerable planning is required for the successful acquisition of ictal scans. Intracranial recordings obtained using long-term implanted intracerebral or subdural electrodes or surgical ECoG can provide accurate localization of an epileptic focus within a region of the brain. However, the techniques are invasive and findings are obtained at considerable cost. In addition, the small number of brain regions that can be sampled can lead to incorrect conclusions regarding the location of the most important epileptic generators, unless a clear hypothesis has been reached before planning electrode placement. 38 Recently, dipole modeling of interictal epileptic activity using scalp EEG data has been shown to correlate well with the generators of these spikes Dipole modeling of median nerve SSEPs has also been shown to identify the location of the primary somatosensory hand area accurately Dipole modeling thus has the potential unique advantage of using the same technique to identify the location of both the somatosensory hand area and the epileptic generator. The use of dipole modeling of sources in three-dimensional space has greatly improved the resolution of EEG. However, the accuracy of the model can at times be difficult to predict. 45 Gross et al, 46 assessed the spatial accuracy of dipole localization in determining both epileptic spikes and the upper limb area by SSEP. They concluded that the calculation of the relative location of spike and SSEP dipoles is a simple noninvasive method of determining the relationship between the primary hand area and an epileptic focus in the central area. The spatial resolution of this technique can be further improved using easily identifiable anatomical landmarks. In our work we aimed to assess the ability of source analysis techniques to improve our knowledge about the possible source generators of epilepsy in a non invasive way. We were not aiming to validate source analysis in this study but only assess its diagnostic yield. The cases were selected for study to meat this aim. They were medically intractable cases that were undergoing extensive non invasive preoperative evaluation. The results of our study are to some extent unique as all cases were partial epilepsy. Most of them were (68%), 20% extra and 12% mixed and extra lobe epilepsy. All cases had a positive neuroimaging finding except one case. The presence of an epileptogenic lesion was important in our cases so that the clinico-electro-anatomical correlation trigone would be complete for localizing the epileptogenic zone. We used MRS in only 4 cases. The later added to our knowledge either by giving more lateralizing information in bilateral cases or by diagnosing MTS in MRI free cases. The method we used for spatial localization of EEG dipole in MRI using the system was first described by Pascua et al in We used standard head model and standard MRI in our technique as the use of realistic head models was of limited added value. 48 We also used dipole localizations of both averaged and individual spikes. Because the use of averaged spikes in dipole localization is still debatable. 49 The Interictal EEG and Ictal EEG was rated by two different rater blinded to each other for more validation of the results. The results of EEG was assessed as regards two things; the lateralizing and the localizing ability of it. By interictal EEG we could lateralize 60% of cases and 68% were by ictal EEG. The difference was statistically insignificant. The interictal EEG could localize 20% as focal and 54% as regional and ictal EEG did not add a lot as 24 % was localized as focal and 64% was localized as regional. Again, the difference was statistically insignificant. 318

17 Ahmed A. Gaber et al. These result would raise a question of the value of video EEG monitoring over the interictal EEG in the presence of this limited spatial resolution. Both have good lateralizing ability with fair regional and minimal focal localization, with no superiority of ictal EEG over interictal EEG. Although, the ictal EEG is still of value in two things, first it verifies the result of interictal EEG and second it is the only objective way to study seizure semiology. However, its value of localization is rather similar to interictal EEG. The use of dipole source analysis in this study proved to be superior in lateralizing and localizing EEG data. The ASA has high significant lateralizing ability than interictal EEG, and highly significant localizing ability over both. It is important to say that all case were localized in focal areas (not regional) with x,y,z landmarks of the dipoles. Four cases had 2 generators and 2 cases had a mid line generators that were hypothalamic hamartoma. At the stage of data integration ASA results were highly complementary to the four landmarks of the epileptogenic zone, the ictal symptomatic zone, irritative zone, pace maker zone and epileptogenic lesion. The concordance between these zones was 32% but with ASA it reached 96% with highly significant difference. The lateralizing ability of data with ASA was not statistically superior. However, only two cases were not lateralized and proved to have mid-line generators. As regards the localization ability before ASA only 4 cases could be localized and 14 cased had regional localization. After ASA all cases could be localized with highly significant difference. This means that ASA contributes significantly to accurate localization with excellent concordance with other clinico electroanatomical data. Our results is going with that of Meckes- Ferber 14 who studied EEG dipole source localisation of interictal spikes acquired during routine clinical video-eeg monitoring and found that data derived from routine clinical inpatient VEM using a routine montage can yield accurate EEG dipole source localization, but significantly more accurate localization is obtained using the surgical montage. Although the results of ASA are favorable, but it should be taken with caution when epilepsy surgery is considered. Not only the localization of the epileptic spike but the timing of the sources of an epileptic spike is important. If these sources are thought of as extended areas of cortical tissue that are simultaneously active, application of dipole models may not be valid. Validation can be done when invasive measurements exist There is also many questions that should be answered. Is the interictal EEG enough for accurate dipole source localization or ictal recording is needed? Is there is a difference in sensitivity of dipole source analysis between lesional and non lesional cases, and between and extra cases?. All these questions need a lot of work in the future. In conclusion advanced source analysis can provide complementary diagnostic yield to the preoperative assessment of intractable epilepsy with more precise anatomical localization of the pace maker however it still needs validation. In our opinion the only validation for source analysis is the invasive EEG recording whether ECoG or stereoeeg. REFERENCES 1. Beghi E, Garattini L, Ricci E, EPICOS Group, et al. Direct cost of medical management of epilepsy among adults in Italy: a prospective cost-of-illness study (EPICOS). Epilepsia 2004; 45: Rosenow F, Luders H. Presurgical evaluation of epilepsy. Brain 2001; 124: Engel J Jr. Surgery for seizures. N Engl J Med 1996; 334: Engel J Jr. Finally, a randomized, controlled trial of epilepsy surgery. N Engl J Med 2001; 345: Wiebe S, Blume WT, Girvin JP, et al. Effectiveness and efficiency of surgery for lobe epilepsy study group. A randomized, controlled trial of surgery for -lobe epilepsy. N Engl J Med 2001; 345: Engel J Jr, Williamson PD, Wieser HG. Mesial lobe epilepsy. In: Engel J Jr, Pedley J, eds. Epilepsy: a comprehensive textbook. New York: Raven Press, 1997:

18 Egypt J. Neurol. Psychiat. Neurosurg. Vol. 43 (1) Jan Engel J Jr, Cascino GD, Shields WD. Surgical remediable syndromes. In: Engel J Jr, Pedley J, eds. Epilepsy: a comprehensive textbook. New York: Raven Press, 1997: Engel J Jr. Principles of epilepsy surgery. In: Shorvon S, Dreifuss F, Fish D, et al, eds. The treatment of epilespy. Oxford: Blackwell Sciences, 1996: Quarato PP, Di Gennaro G, Mascia A, et al. Temporal lobe epilepsy surgery: different surgical strategies after a non-invasive diagnostic protocol. J Neurol Neurosurg Psychiatry 2005; 76: Bancaud J, Talairach J, Bonis A, et al. La Ste re oe lectroence phalographie dans l e pilepsie. Paris: Masson, Quated from Quarato PP, Di Gennaro G, Mascia A, et al. Temporal lobe epilepsy surgery: different surgical strategies after a non-invasive diagnostic protocol. J Neurol Neurosurg Psychiatry 2005; 76: Munari C, Tassi L, Lo Russo G, et al. La terapia chirurgica nelle epilessie parziali: schematizzazione metodologica o approccio personalizzato. Bollettino della Lega Italiana Contro L Epilessia 1990;70/1:61 7. Quated from Quarato PP, Di Gennaro G, Mascia A, et al. Temporal lobe epilepsy surgery: different surgical strategies after a non-invasive diagnostic protocol. J Neurol Neurosurg Psychiatry 2005; 76: Bartolomei F, Guye M, Gavaret M, et al. The presurgical evaluation of epilepsies. Rev Neurol (Paris) 2002; 158(5 Pt 2): 4S Chauvel P, Halgren E. Les illusions et le hallucinations complexes dans le crises du lobe. In: Crises epileptiques et epilepsies du lobe, Vol. I. Gentilly: Documentation Medicale Labaz, 1991: Quated from Quarato PP, Di Gennaro G, Mascia A, et al. Temporal lobe epilepsy surgery: different surgical strategies after a non-invasive diagnostic protocol. J Neurol Neurosurg Psychiatry 2005; 76: Meckes-Ferber S; Roten A; Kilpatrick C; O'Brien TJ. EEG dipole source localisation of interictal spikes acquired during routine clinical video- EEG monitoring. Clin Neurophysiol 2004 Dec; 115(12): Fuchs M; Ford MR; Sands S; Lew HL. Overview of dipole source localization. Phys Med Rehabil Clin N Am 2004 Feb; 15(1): Luders HO, Awad IA. Conceptual considerations. In: Luders HO, ed. Epilepsy surgery. New York: Raven Press, 1992: Palmini A, Gloor P. The localizing value of auras in partial epilepsies. Neurology 1992; 42: Henkel A, Noachtar S, Pfander M, et al. The localizing value of the abdominal aura and its evolution. Neurology 2002; 58: Di Gennaro G, Quarato PP, Onorati P, et al. Localizing significance of intermittent rhythmic delta activity (TIRDA) in drug-resistant focal epilepsy. Clin Neurophysiol 2003;114: Clusmann H, Schramm J, Kral T, et al. Prognostic factors and outcome after different types of resection for lobe epilepsy. J Neurosurg 2002; 97: Garcia PA, Laxer KD, Barbaro NM, et al. Prognostic value of qualitative magnetic resonance imaging hippocampal abnormalities in patients undergoing lobectomy for medically refractory seizures. Epilepsia 1994; 35: Eskola H.Utilization of MRI Information in EEG Studies. International Journal of Bioelectromagnetism, IJBEM 1999, 1(1), Mosher JC, Leahy RM, Recursive MUSIC: a framework for EEG and MEG source localization. IEEE Trans Biomed Eng 1998; 45(11): Malmivuo J, Plonsey R, Bioelectromagnetism - Principles and Applications of Bioelectric and Biomagnetic Fields. Oxford University Press, New York, Brazier M, A study of the electric field at the surface of the head. Electroencephalogr Clin Neurophysiol 1949; 2(Suppl 1): Berg P, Scherg M, A fast method for forward computation of multiple-shell spherical head models. Electroencephalogr Clin Neurophysiol 1994; 90(1): Laarne P, Eskola H, Hyttinen J, Suihko V, Malmivuo J, Validation of a detailed computer model for the electric fields in the brain. J Med Eng Technol 1995; 19(2-3): Zanow F, Peters MJ, Individually shaped volume conductor models of the head in EEG source localisation. Med Biol Eng Comput 1995; 33(4): Lemieux L, McBride A, Hand J, Calculation of electrical potentials on the surface of a realistic head model by finite differences. Phys Med Biol 1996; 41(7):

19 Ahmed A. Gaber et al. 30. Haueisen J, Ramon C, Eiselt M, Brauer H, Nowak H, Influence of tissue resistivities on neuromagnetic fields and electric potentials studied with a finite element model of the head. IEEE Trans Biomed Eng 1997; 44(8): Fuchs M, Drenckhahn R, Wischmann HA, Wagner M, An improved boundary element method for realistic volume-conductor modeling. IEEE Trans Biomed Eng 1998; 45(8): Towle VL, Bolanos J, Suarez D, Tan K, Grzeszczuk R, Levin DN, Cakmur R, Frank SA, Spire JP, The spatial location of EEG electrodes: locating the best-fitting sphere relative to cortical anatomy. Electroencephalogr Clin Neurophysiol 1993; 86(1): Gaber A, Etribi M A, Zaki A et al: Assessment of novel strategies in the management of intractable epilepsy. MD Thesis, Ain Shams University Ebersole JS, Pacia SV. Localization of lobe foci by ictal EEG patterns. Epilepsia 1996; 37: Pacia SV, Ebersole JS. Intracranial EEG substrates of scalp ictal patterns from lobe foci. Epilepsia 1997;38: Ebersole JS. Localization of epileptic foci in lobe epilepsy. Electroencephalogr Clin Neurophysiol 1997; 103: Luders HO: The supplementary sensorimotor area: an overview. Adv Neurol 1996; 70 : Gloor P: Neuronal generators and the problem of localization in electroencephalography: application of volume conductor theory to electroencephalography. J Clin Neurophysiol 1985; 2: Boon P, D Have M, Vandekerckhove T, et al: Dipole modeling and intracranial EEG recording: correlation between dipole and ictal onset zone. Acta Neurochir 1997; 139: Merlet I, Garcia-Larrea L, Ryvlin P, et al: Topographical reliability of mesio- sources of interictal spikes in lobe epilepsy. Electroencephalogr Clin Neurophysiol 1998; 107: Merlet I, Gotman J: Reliability of dipole models of epileptic spikes. Clin Neurophysiol 1999; 110: Buchner H, Adams L, Knepper A, et al: Preoperative localization of the central sulcus by dipole source analysis of early somatosensory evoked potentials and three-dimensional magnetic resonance imaging. J Neurosurg 1994; 80: Buchner H, Adams L, Müller A, et al: Somatotopy of human hand somatosensory cortex revealed by dipole source analysis of early somatosensory evoked potentials and 3D-NMR tomography. Electroencephalogr Clin Neurophysiol 1995; 96: Mine S, Oka N, Yamaura A, et al: Presurgical functional localization of primary somatosensory cortex by dipole tracing method of scalp-skullbrain head model applied to somatosensory evoked potential. Electroencephalogr Clin Neurophysiol 1998; 108: van den Broek SP, Reinders F, Donderwinkel M, et al: Volume conduction effects in EEG and MEG. Electroencephalogr Clin Neurophysiol 1998; 106: Gross D W, Merlet I, Boling W and Gotman J: Relationships between the epileptic focus and hand area in central epilepsy: combining dipole models and anatomical landmarks. J Neurosurg 2000; 92: Pascau J, Desco M, Rojo P, et al.: Spatial localisation of EEG dipoles in MRI using the International System anatomical references. Proc. of First Int'l Workshop on Image and Signal Processing and Analysis Vanrumste B, Van Hoey G, Van de Walle R, et al.: The Realistic Versus The Spherical Head Model In Eeg Dipole Source Analysis In The Presence Of Noise. Proceedings 23rd Annual Conference IEEE/EMBS Oct , 2001, Istanbul, TURKEY 49. Chitoku S, Otsubo H, Ichimura T, et al.: Characteristics of dipoles in clustered individual spikes and averaged spikes. Brain Dev 2003 Jan; 25(1): Ebersole J: EEG source modeling. The last word. Journal of Clinical Neurophysiology 1999; 16: Greensite F and Huiskamp G: Temporally unconstrained space-time treatment of linear formulations of the inverse problem of electroencephalography. Annals of Biomedical Engineering 2000; 28: