Source localization of scalpEEG interictal spikes ... - Wiley Online Library

5 downloads 833 Views 945KB Size Report
Jun 6, 2008 - Discussion: The accuracy of scalp-EEG IIS source localization is ..... each electrode contact during video SEEG monitoring were reviewed and ...
Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

FULL-LENGTH ORIGINAL RESEARCH

Source localization of scalp-EEG interictal spikes in posterior cortex epilepsies investigated by HR-EEG and SEEG ∗

yzMartine Gavaret, ∗ yzAgne`s Tre´buchon, ∗ yzFabrice Bartolomei, ∗ yPatrick Marquis, ∗ yzAileen McGonigal, x{Fabrice Wendling, ∗ y#Jean Regis, ∗ yJean-Michel Badier, and ∗ yzPatrick Chauvel

∗ INSERM UMR 751, Laboratoire de Neurophysiologie et Neuropsychologie, Marseille, France; yAix Marseille Universite´, Faculte´ de Me´decine, Marseille, France; zAssistance Publique Hoˆpitaux de Marseille, Hoˆpital de la Timone, Service de Neurophysiologie Clinique, Marseille, France; xINSERM, U642, Rennes, France; {Universite´ de Rennes 1, LTSI, Rennes, France; #Assistance Publique Hoˆpitaux de Marseille, Hoˆpital de la Timone, Service de Neurochirurgie Fonctionnelle, Marseille, France

SUMMARY Purpose: To determine the validity of scalp-electroencephalography (EEG)-interictal spike (IIS) source localization in posterior cortex epilepsies (PCE). Methods: Eleven patients with drug-resistant PCE were studied with high-resolution EEG (HR-EEG) and stereoelectroencephalography (SEEG). Sixtyfour scalp channels, a realistic head model, and different algorithms [multiple signal classification (MUSIC) and equivalent current dipoles] were used. Results were compared to intracerebral SEEG recordings. For SEEG, a semiautomatic detection of intracerebral IIS was used, allowing a classification of intracerebral IIS into one of three groups: Medial, lateral, and mediolateral. Results: In the medial group (two patients), scalpEEG IIS were usually absent for one patient whereas for the other, scalp-EEG was misleading. Indeed, scalp-EEG IIS had a posterior projection,

Posterior cortex epilepsies (PCE; originating from the occipital, parietal, or occipital border of the temporal lobe) are relatively rare (Sveinbjornsdottir & Duncan, 1993; Boesebeck et al., 2002) and challenging in the context of Accepted June 6, 2008; Early View publication August 20, 2008. Address correspodence to Martine Gavaret, INSERM U751 & Service de Neurophysiologie Clinique, CHU Timone, 264 rue Saint Pierre, 13005 Marseille, France. E-mail: [email protected] Wiley Periodicals, Inc. ª 2008 International League Against Epilepsy

predominantly contralateral to the source. In the lateral group (two patients), scalp-EEG IIS were subcontinuous and accurately localized. In the mediolateral group (seven patients), intracerebral interictal distribution was complex and bilateral for four of seven patients. Source localizations were able to determine only a part, whether lateral or medial, of the intracerebral interictal distribution. Discussion: The accuracy of scalp-EEG IIS source localization is dependant on the type of intracerebral interictal distribution. In the most frequent type of PCE, patients proved to have a complex interictal distribution between both medial and lateral cortices, and source localizations always underestimated intracerebral IIS. In cases where intracranial sources were quite focal, surface EEG sources were localized with accuracy, even in medial occipital lobe structures. KEY WORDS: Posterior cortex epilepsy, Occipital epilepsy, Scalp-EEG interictal spikes, Source localization, SEEG.

epilepsy surgery, being characterized by less satisfying postoperative results than temporal lobe epilepsies (Blume et al., 1991; Bautista et al., 1999; Boesebeck et al., 2002; Barba et al., 2005). Since there are no clear anatomic distinctions between occipital, parietal, and temporal lobes, their classification is particularly difficult. A stereoelectroencephalography (SEEG) study of occipital epilepsies demonstrated the frequent propagation of interictal and ictal activities from occipital foci to temporal or parietal lobes or both (Bancaud, 1969). Temporal

276

277 Source Localization in Posterior Epilepsy spread was usually observed when the site of origin was below the calcarine fissure, while parietal spread was generally observed when the site of origin was above the calcarine fissure (Bancaud, 1969; Ludwig & Ajmone Marsan, 1975). Similarly, parietal or temporal epilepsies can propagate to the occipital lobe. Some authors have thus proposed the term of PCE, encompassing a group of epilepsies originating from the occipital, parietal, or occipital border of the temporal lobe, or from any combination of these regions (Blume et al., 1991; Sveinbjornsdottir & Duncan, 1993; Dalmagro et al., 2005). Epilepsies originating from these regions are thus probably better analyzed when grouped together (Dalmagro et al., 2005). Scalp-electroencephalography (EEG) interictal spikes (IIS) are present in 80%–90% of PCE (Blume et al., 1991; Boesebeck et al., 2002). Scalp-EEG IIS can be characterized by a widespread reflection over one posterior quadrant (Westmoreland, 1998; Daniel et al., 2007) or a posterior bilateral distribution (Salanova et al., 1992; Blume et al., 2005; Lee et al., 2005). Scalp-EEG IIS can also be present over the temporal or centrotemporal regions (Olivier et al., 1982; Palmini et al., 1999; Boesebeck et al., 2002; Blume et al., 2005). Scalp EEG may show other abnormalities, including slowing, asymmetry, or attenuation of activity over one or both occipital regions (Westmoreland, 1998). Scalp EEG in PCE is considered to be often misleading (Williamson et al., 1992; Palmini et al., 1993; Sveinbjornsdottir & Duncan, 1993), and some authors have found that a low yield of localizing information provided by scalp-EEG is thus obtained (Palmini et al., 1993). There are few studies of EEG source imaging in PCE in the literature. Several EEG source imaging studies, concerning various types of epilepsies, included some PCE cases (Boon et al., 1997; Huppertz et al., 2001; Merlet & Gotman, 2001; Lantz et al., 2003; Michel et al., 2004a, Sperli et al., 2006, and five cases of idiopathic occipital epilepsy in Leal et al., 2007). In these studies, EEG source localization results were mostly considered as accurate, down to a sublobar level, but validation was indirect, being performed by comparison with other imaging techniques and postoperative outcome. Few previous studies have been validated by intracranial recordings (Boon et al., 1997; Merlet & Gotman, 2001). In these two studies, the number of PCE cases is too small to draw definitive conclusions. Moreover, some of these studies include surface signal analysis without source localization (Boon et al., 1997). It is thus unknown if interictal sources are just difficult to localize from scalp-EEG or if scalp-EEG IIS effectively reflect widespread activity in depth recording. As in other types of epilepsies, a comprehensive study of interictal activities is essential in presurgical investigations. In addition, it has been demonstrated that a better surgical outcome in PCE is correlated to the presence of focal IIS included in surgical resection (Bautista et al., 1999).

The orientation of the generator, its extent (Tao et al., 2007), and the distance separating the generator from surface electrodes are essential parameters to understand problems of detection encountered with scalp-EEG. It has been shown that medial frontal IIS are detected in scalpEEG and accurately localized (Gavaret et al., 2006), whereas medial temporal IIS are not detected in scalpEEG (Gavaret et al., 2004) unless using averaged EEG time locked to the IIS peak in foramen ovale electrode recordings (Zumsteg et al., 2005). We thus face the same questions in PCE, particularly concerning the detection of depth medial posterior sources. We have thus prospectively studied scalp-EEG IIS source localization in 11 drug-resistant PCE patients, in the context of presurgical assessment. Each patient subsequently underwent intracerebral recording using SEEG. Source localization results were then compared to the intracerebral IIS distribution, determined by the SEEG. In this work, we aimed at determining the accuracy of scalp-EEG IIS source localization according to their intracerebral distribution.

Patients and Methods Patients Eleven drug-resistant PCE patients underwent highresolution EEG (HR-EEG) and subsequently SEEG during presurgical investigation. Informed consent was obtained from each patient, and this study was approved by the ethical committee (CCPPRB) of Marseille. These patients had a first-step evaluation including careful history taking, neurological examination, video EEG recording of seizures, structural magnetic resonance imaging (MRI), interictal € ictal single photon computerized tomography (SPECT), positron emission tomography (PET), and neuropsychological evaluation. For each patient, because these noninvasive investigations alone did not permit definition of the epileptogenic zone, subsequent SEEG was indicated to define the contribution of distinct posterior areas to the epileptogenic zone. Patients’ characteristics are detailed in Table 1. HR-EEG data Recording According to our previous studies (Gavaret et al., 2004; Gavaret et al., 2006), HR-EEG data were recorded using 64 scalp electrodes, referenced to a mastoidal electrode contralateral to habitual predominant scalp-EEG IIS (1010 system, Quickcap; Compumedics Neuroscan, El Paso, TX, U.S.A.). Impedances were all below 5 kX. The signal was recorded during four sessions of 10 min at a 1000 Hz sampling rate. The activity was digitally filtered with a bandpass 0.15–200 Hz (SynAmps; Compumedics Neuroscan, Charlotte, NC, U.S.A.). During sessions, the patient was relaxed with eyes closed, awake, in a sound attenuated Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

Gender

Age

Initial ictal symptoms

Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

14

Subcontinuous

Subcontinuous

Abundant Present Subcontinuous Absent

Present Subcontinuous



FCD Gliosis

— — FCD Gliosis Gliosis FCD —

R supracalcarine FCD

R lateral temporooccipital FCD Normal

R lateral occipital lesion R medial occipital DNET Previous R occipital epilepsy surgery R medial parietooccipital lesion Normal Normal Normal

Subcontinuous

Subcontinuous

Absent

Occurrence

Gliosis

Histologic diagnosis

L occipital surgical scar

Structural MRI

Bilateral R > L

Bilateral R > L

Bilateral L > R

Bilateral R > L Bilateral 2 distinct types Bilateral 2 distinct types —

Bilateral R > L

FT8, F10, FT10

P4, PO4, CP4

80–110

60–70

80–90



— P3, CPz, Pz

70–80

70–90 80–90

60–70

60–70

60–70



Duration

FT10, T,8 TP8, P3, P9, P7

P4, PO8, PO4 CPz, Pz, P8, P10, PO8

PO4, P4, PO8

TP8, PO10, P8

Pz, Poz, Oz, PO3

Bilateral R ¼ L

Bilateral R > L



Dominant channels



Topography

Scalp-EEG IIS

>3

>3

>3



>3

>3 2

>3

>2

>3



SNR

Patients are classified according to their intracerebral interictal distribution: Medial, lateral, and mediolateral. Patient characteristics are indicated as follows: Gender (two males/ nine females), age at the time of the SEEG (14 to 48 years), initial ictal symptoms observed during video EEG, structural MRI data, histologic diagnosis, scalp-EEG IIS characteristics: Occurrence, topography, dominant channels, duration (ms), and SNR. FCD, focal cortical dysplasia. FCD indicated in structural MRI data means MRI suggestive of focal cortical dysplasia (gyral thickening with abnormal MRI signal), usually T2 prolongation in the underlying white matter, according to the criteria of Kuzniecky & Barkovich, 2001. inf, inferior; IIS, interictal spike; L, left; R, right; SNR, signal-to-noise ratio.

F

Loss of contact, eyes and head deviation to the L Mediolateral posterior intracerebral IIS (posterior M/L index: 0.5–4) 5-PS F 24 Visual illusions, loss of contact, laughter 6-RC F 30 Loss of contact, automatic movements with both legs, laughter 7-BL M 28 Bilateral blinking, eyes and head deviation to the L, smile 8-FC F 35 Complex visual hallucinations, loss of contact, head deviation to the L 9-ZC F 26 Visual hallucinations and illusions on the R, eyes and head deviation to the L 10-ML F 23 Elementary visual hallucinations, loss of contact 11-VC F 48 Elementary visual hallucinations, loss of contact, head deviation

4-DS

Medial posterior intracerebral IIS (posterior M/L index > 4) 1-LF M 19 Flashing white lights in the R inf quadrant, generalization 2-LN F 41 Amaurosis or palinopsia, loss of contact, enuresis Lateral posterior intracerebral IIS (posterior M/L index < 0.5) 3-BA F 17 Visual illusions, ocular jerks, facial flushing

Patients

Table 1. Posterior epilepsy patient and IIS characteristics

278 M. Gavaret et al.

279 Source Localization in Posterior Epilepsy and electrically shielded room. Activation procedures such as hyperventilation, withdrawal of antiepileptic drugs, or drug-induced activation were not used. Digitization of electrode positions Before EEG recording, positions of all electrodes and of three marking points (nasion, right, and left tragi) were registered with a three-dimensional (3D) digitizer system (3Space Fastrak; Polhemus, Colchester, VT, U.S.A.). These three marking points were used to define a spatial reference. The patient’s forehead and nasal surface were also registered by collecting approximately 600–700 points while moving the digitizer across the head contour. Scalp-EEG IIS detection, selection, and analysis On both bipolar and referential montages, scalp-EEG IIS (spikes, spike-waves, or sharp waves) were identified and selected according to the International Federation of Societies for Electroencephalography and Clinical Neurophysiology (IFSECN) criteria (Chatrian et al., 1974). Scalp-EEG IIS occurrence, without withdrawal of antiepileptic drugs, was evaluated as (1) absent; (2) rare, indicating that less than five scalp-EEG IIS were recorded during the HR-EEG (40 min); (3) present, being used when 5 to 20 scalp-EEG IIS were recorded; (4) abundant, when more than 20 scalpEEG IIS were recorded; and (5) subcontinuous, when they were organized in subcontinuous bursts (Table 1). Scalp-EEG IIS were marked at the time of their highest amplitude (using EEGFocus; MEGIS Software, Grfelfing, Germany). Epochs of 1000 ms were extracted (500 ms before and after the highest amplitude of the event) and transferred to Advanced Signal Analysis (ASA) software (ANT Software, Enschede, The Netherlands). Channels that recorded scalp-EEG IIS with maximal amplitude (dominant channels) were determined in ASA using a common average reference. The duration of IIS was defined by the duration of several single spikes without the following slow waves, seen at the dominant electrode with a common average reference. Signal-to-noise ratio (SNR) was defined by scalp-EEG IIS maximal amplitude divided by background activity maximal amplitude. SNR was calculated on single IIS, the 64 channels being superimposed in ASA. Realistic head model The 3D MRI was based on a T1-weighted sequence, with pixel size of 1.25 mm2 and slice thickness of 1.25 mm. We identified on the 3D MRI the same three marking points (nasion, right, and left tragi). We also defined the same spatial reference for neurophysiological and structural data. The realistic head model was based on a boundary element method (BEM), which describes the individual surfaces by triangulation, with about 1,700–2,000 nodes per model. The segmentation process, identification of the three compartments of isoconductivity (intracranial space, skull, and

scalp) was performed with ASA. We used an automatic 3D segmentation, the skull being the dilatation (·3) of the intracranial space. Triangulation of the surfaces was performed automatically. Based on these triangulation data, a realistic EEG transfer matrix was calculated (Zanow & Peters, 1995). This realistic head model took into consideration individual differences in head volume. Electrical impedance tomography (GonÅalves et al., 2000) was not performed in this study. Different specific conductivities were attributed to each volume according to Geddes and Baker (1967) (0.33 Siemens/meter for intracranial space and scalp, 0.0042 Siemens/meter for skull). Fusion of morphological and neurophysiological data Electrode positions, the patient’s head contour, and 3D MRI all had the same spatial reference determined by the three marking points previously described. The positions of the electrodes and patient’s head contour were then superimposed upon the 3D MRI to exactly match the two surfaces and adjusted manually if necessary. Multiple Signal Classification To get an estimation of the contribution of generators spatially distributed within the intracranial volume, we used the so-called multiple signal classification (MUSIC) method. This method was adapted from a scanning technique used in radar technology to isolate signal from noise (Mosher et al., 1999). The method is based on an eigenvalue decomposition of the data to identify the underlying components (the signal space) in the time series data (Mosher et al., 1999; Michel et al., 2004b). The whole brain volume is scanned for those source locations that contribute to the signal space. Temporal window of analysis corresponded to the IIS duration (60–110 ms). For each patient, MUSIC was carried out on at least 10 single IIS by means of ASA software. The number of components explaining at least 95% of the signal was selected for the analysis. In the intracranial volume, a metric was built, corresponding to n equidistant (8 mm) equivalent current dipoles. Single unconstrained equivalent current dipole modeling Single unconstrained equivalent current dipole modeling was carried out on at least 10 single scalp-EEG IIS. This analysis was first performed over the time course of single scalp-EEG IIS with a moving dipole approach (i.e., calculation of a new dipole localization, orientation, and amplitude for each millisecond). This allows determination of whether a particular area can be identified as the origin of the activity when the localization of the equivalent current dipole remains stable in the same area for sufficient time (20–30 ms). Apparent movement of the dipole between distant structures will reveal an inadequate source model or propagation of the activity (unstable localization can also appear in the case of insufficient SNR). For a moving dipole, we retained its position where Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

280 M. Gavaret et al. correlation (percentage of the EEG data that could be explained by the model) was maximal (generally corresponding to the peak of scalp-EEG IIS). A spatiotemporal model (single unconstrained fixed dipole/single unconstrained rotating dipole), i.e., same localization within the time window, was used later on a time window corresponding to scalp-EEG IIS duration. SEEG data SEEG recording SEEG was performed using multicontact depth electrodes (diameter 0.8 mm, 10 to 15 contacts, each 2 mm long with an intercontact distance of 1.5 mm) implanted intracerebrally according to Talairach’s stereotactic method (Talairach et al., 1974; McGonigal et al., 2007). Precise localization of the target cortical structures was based on surface rendering of the 3D T1-weighted sequence, on 3D MRI reconstruction, and on peroperative telemetric angiography. In these posterior epilepsy patients, 7 to 16 electrodes were orthogonally implanted (Table 2), providing 70 to 128 points of measurement in each case, and consequently an extended electrophysiological sampling of the brain areas of interest. The cortical targets for the positioning of electrodes were determined by the clinical, neurophysiological, and anatomical characteristics of each patient. The exact position of each depth electrode contact was ascertained postoperatively, using the immediate postoperative computed tomography (CT) scan and the MRI performed following electrodes removal (3D T1-weighted images and T2-weighted coronal images). SEEG analysis For each orthogonally implanted depth electrode, EEG recordings were obtained from 15 contacts. The external contacts (numbered 10–15) recorded lateral infolded cortex and gyral crowns. Respectively, the internal contacts (numbered 1–5) recorded depth medial cortex. The intermediate part of each depth electrode (contacts 6–9) was generally localized in the white matter. The sulcal pattern of the occipital lobe is known to vary considerably between individuals (Ono et al., 1990), and the overall position of posterior areas described here was defined according to Braak (1977), Clarke and Miklossy (1990), and Clarke (1994). For each patient, we have used the same terminology to determinate the anatomical position of depth electrodes. A typical SEEG scheme is depicted in Fig. 1. In the occipital lobe, dorsomedial areas (above the calcarine fissure) of the occipital lobe (cuneus) were explored by medial contacts of the depth electrode called Cu. Ventromedial areas (below the calcarine fissure) of the occipital lobe (lingual gyrus, occipitotemporal junction) were respectively explored by the depth electrodes called GL and OT (medial contacts). Dorsolateral areas of the occipital lobe were explored by the depth electrode Cu (lateral contacts). Ventrolateral areas of the occipital lobe Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

and the lateral occipitotemporal junction were respectively investigated by the lateral contacts of electrodes GL and OT. In the parietal lobe, different areas could be explored by SEEG. The depth electrode called GC explored the posterior part of the supramarginalis gyrus (lateral contacts) and the isthmus of the cingulate gyrus (medial contacts). An electrode PS was localized in the dorsal part of the parietal lobe. An electrode PI was localized in the ventral part of the parietal lobe. An electrode OP investigated the parietal operculum (lateral contacts) and the posterior part of the insula (medial contacts). In the temporal lobe, different depth electrodes (medial/lateral contacts) could be positioned as follows: A (amygdala/middle temporal gyrus), B (hippocampal head/middle temporal gyrus), C (hippocampal tail/middle temporal gyrus), T (insula/superior temporal gyrus), TB (temporobasal areas), and TP (temporal pole). In the frontal lobe, for some patients, an electrode SA (supplementary motor area) was positioned. For 7 of 11 patients, SEEG implantation was bilateral (Table 2). Signals were recorded on a 128-channel Deltamed(tm) system. They were sampled at 256 or 512 Hz and recorded on a hard disk (16 bits/sample) using no digital filter. The only filter present in the acquisition procedure was a hardware analog high-pass filter (cut-off frequency equal to 0.16 Hz) used to remove very slow variations that sometimes contaminate the baseline. SEEG signals recorded by each electrode contact during video SEEG monitoring were reviewed and analyzed to determine whether or not depth IIS were present in the corresponding brain area. Detection and analysis of IIS In order to detect intracerebral IIS, we used a previously described algorithm (Bourien et al., 2004, 2005). This method stems from the fact that IIS generally include a sharp component corresponding to a transient wave of high amplitude and short duration compared to background activity. This component is characterized by a specific signature in the time-frequency plane, i.e., an increase of energy in higher frequency bands (typically from 20 to 40 Hz). Based on this observation, the detection method included two main steps. In the first stage, each EEG signal was decomposed on a wavelet filter bank. The mean value q(t) of the squared modulus of filter outputs was computed at each sample time (Senhadji et al., 1995). By construction, this random quantity q(t) exhibits high mean value during IIS duration and low mean value during background EEG. In the second stage, a Page-Hinkley algorithm (Basseville & Nikiforov, 1993) was used to automatically estimate time instants corresponding to abrupt changes of q(t), each abrupt change corresponding to the occurrence of an IIS in the analyzed monochannel SEEG signal. For each patient, we performed intracerebral IIS detection on selected representative SEEG signals recorded from medial and lateral contacts of each depth electrode. A 1-h period of continuous interictal SEEG was selected

Posterior M/L index

N Elc

Cu

GL

OT X/o X/o ) o/o ) XX/o ) XX/o o/X ) )

) o/XX ) ) ) o/o ) ) )

o/X X/XX X/o XX/o o/o XX/X X/o o/o X/o

PI

) X/o

PS

X/o XX/o

GC

) ) o/o o/o o/o ) )

) o/o

o/o )

OP

) ) ) o/o ) ) XX/X

) )

) )

A

X/o o/o o/o ) o/o X/o )

o/X X/o

) )

B

) X/X o/o o/o X/XX X/o X/o

o/X )

) X/o

C

o/o o/o o/o o/o o/o ) o/o

) )

) )

T

Temporal lobe

X/o o/o o/o ) ) X/o )

o/X X/X

) )

TB

SEEG: medial/lateral intracerebral electrodes contacts Parietal lobe

) ) ) ) ) ) XX/XX

) )

) )

TP

) ) o/o o/o o/o ) o/o

) o/o

) o/o

SA

Frontal lobe

XX/X X/o o/X ) ) o/o X/XX

) )

) o/o

OL

) X/o ) ) o/o ) o/o

) o/XX

) XX/o

PL

Contralateral

X/o X/o o/X ) ) o/o X/X

) o/o

) o/o

TL

Semiautomatic detection of intracerebral IIS was performed on occipital, parietal posterior, and temporal posterior depth electrodes. Posterior M/L index value was obtained by dividing the number of intracerebral IIS detected in medial contacts by the number of intracerebral IIS extracted from lateral contacts. N Elc indicates the number of depth electrodes implanted for each patient. OL, occipital lobe; PL, parietal lobe; TL, temporal lobe; FL, frontal lobe. Contralateral means in the contralateral hemisphere. Depth electrode recorded from specific regions as follows (medial/lateral contacts): Cu, cuneus/dorsolateral areas of the occipital lobe; GL, lingual gyrus/ventrolateral areas of the occipital lobe; OT, occipitotemporal junction; GC, isthmus of the cingulate gyrus/supramarginalis gyrus; PS, dorsal part of the parietal lobe; PI, ventral part of the parietal lobe; OP, posterior part of the insula/ parietal operculum; A, amygdala/middle temporal gyrus; B, hippocampus head/middle temporal gyrus; C, hippocampus tail/middle temporal gyrus; T, insula/superior temporal gyrus; TB, temporobasal areas; TP, temporal pole; SA, supplementary motor area. XX, presence of intracerebral IIS; X, rare intracerebral IIS; o, absence of intracerebral IIS; ), absence of depth electrode.

Medial posterior intracerebral IIS (posterior M/L index > 4) 1-LF 6.94 7 XX/o X/o o/o 2-LN 4.20 12 o/o o/o o/o Lateral posterior intracerebral IIS (posterior M/L index < 0.5) 3-BA 0.31 8 X/XX o/X o/X 4-DS 0.44 11 o/XX X/XX ) Mediolateral posterior intracerebral IIS (posterior M/L index: 0.5–4) 5-PS 3.02 12 X/X XX/X o/X 6-RC 2.74 11 ) X/X ) 7-BL 1.39 13 o/XX X/XX X/o 8-FC 1.33 12 X/X o/X ) 9-ZC 0.86 11 X/XX X/XX ) 10-ML 0.81 10 X/X XX/XX X/X 11-VC 0.50 16 ) o/o XX/XX

Patients

Occipital lobe

Table 2. Intracerebral interictal distribution determined by SEEG

281

Source Localization in Posterior Epilepsy

Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

282 M. Gavaret et al.

Figure 1. Scalp EEG, source localization and intracerebral IIS for Patient 2 (medial posterior group). (A) Scalp-EEG IIS were subcontinuously recorded over both parietooccipital regions, their amplitude being maximal on channels Pz, POz, Oz, PO3 (monopolar montage, common average reference). We have selected, for source localization, scalp-EEG interictal spikes at the beginning of rhythmic bursts (first scalp-EEG spike surrounded), in order to avoid as possible propagated interictal activities. (B) Amplitude cartography (using Focus software), during the first scalp-EEG IIS surrounded in panel A, showed a maximal amplitude on left parietooccipital regions. (C) The same single IIS, 64 channels superimposed. Temporal window of analysis lies between the two vertical lines. (D) Source localization using MUSIC demonstrated a posterior medial source, bilateral but predominant in the medial part of the right occipitoparietal junction, thus contralateral to scalp-EEG IIS. (E) SEEG schema superimposed upon patient’s 3D MRI. SEEG was performed using 12-depth electrodes. Each depth electrode recorded from specific regions as follows (medial/lateral contacts) in the right hemisphere: Cu (cuneus/angular gyrus), GC (isthmus of the cingulate gyrus at the level of the anterior calcarine fissure/horizontal ramus of the superior temporal sulcus), GL (lingual gyrus/lateral occipitotemporal junction), OT (occipitotemporal junction), PS (dorsal part of the parietal lobe), PI (ventral part of the parietal lobe), C (hippocampal tail/posterior middle temporal gyrus), SA (supplementary motor area); and respectively in the left hemisphere with electrodes Cu¢, GC¢, GL¢, and PS¢. Intracerebral IIS were recorded bilaterally by medial contacts (in yellow) of depth electrodes GC and GC¢. A slightly increased signal in this T2-weighted fluid-attenuated inversion recovery (FLAIR) sequence was detected in the subcortical white matter, above the right calcarine fissure. Epilepsia ILAE

with precautions regarding seizure occurrence, state of vigilance of the patient, residual effect of anesthesia, and antiepileptic drug withdrawal. SEEG period was thus selected during the second day of SEEG exploration (same Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

time following anesthesia for all patients and beginning of antiepileptic drug withdrawal), when the patient was awake. All selected periods were temporally distant from seizure periods (at least 2 h).

283 Source Localization in Posterior Epilepsy Additionally, for each patient, we then calculated a posterior mediolateral (M/L) index. For posterior depth electrodes (occipital, parietal posterior, and temporal posterior), the index value was obtained by dividing the number of intracerebral IIS detected in medial contacts by the number of intracerebral IIS extracted from lateral contacts (Table 2).

Results Patient classification according to intracerebral IIS characteristics For each patient, interictal source localizations were compared with intracerebral interictal distribution to determine the accuracy of HR-EEG in PCE. We have chosen to present the results according to the intracerebral interictal distribution. Visual analysis was completed by a quantitative analysis of the intracerebral data, including semiautomatic detection of intracerebral IIS and calculation of a posterior M/L index. This allowed classification of patients into one of three groups: Medial, lateral, or mediolateral. Two patients were included in the medial group (Patients 1 and 2; Table 2). They were characterized by intracerebral IIS predominant in medial posterior structures, with a mean frequency of 290/h in medial posterior structures versus 65/ h in lateral ones and a posterior M/L index superior to 4. Patients 3 and 4 were classified in the lateral group, being characterized by intracerebral IIS predominant in lateral posterior structures, with a mean IIS frequency of 90/h in medial posterior structures versus 215/h in lateral ones. Their posterior M/L index was inferior to 0.5 (Table 2). Most PCE patients (Patient 5–11) were characterized by intracerebral IIS distributed between medial and lateral structures without clear predominance (Table 2 and Fig. 2). Their mean IIS frequency was 306/h in medial posterior structures and 236/h in lateral ones, with posterior M/L index comprised between 0.5 and 4. These seven patients were classified in the mediolateral group. Morphologic characteristics of scalp-EEG IIS Scalp-EEG IIS were found to be abundant in these patients. Morphologic characteristics of scalp-EEG IIS are detailed in Table 1. Scalp-EEG IIS were recorded for 9 of 11 patients and were characterized by a bilateral projection. These were subcontinuous for 6 of 11 patients. Two distinct types of scalp-EEG IIS were individualized for two mediolateral patients. These scalp-EEG IIS subtypes were selected separately for source localization analysis (Patients 6 and 7; Table 1). Scalp-EEG IIS were characterized by their high amplitude, objectified by their SNR, equal to or higher than 2 for all patients who presented scalp-EEG IIS (Table 1). Two patients (Patient 1 in the medial group and Patient 8 in the mediolateral group) had no scalp-EEG IIS during HR-EEG. The case of Patient 1 was characterized by usual absence of scalp-EEG IIS.

Source localization Source localization results are detailed in Table 3. Source localization for medial group The two patients included in this group (Patients 1 and 2; Table 1) had different scalp-EEG characteristics. Indeed, Patient 1 was characterized by usual absence of scalp-EEG IIS, whereas Patient 2 had subcontinuous scalp-EEG IIS (Table 1). For Patient 2, scalp-EEG IIS were posterior and bilateral with maximal amplitude on midline and on left posterior channels (dominant channels Pz, POz, Oz, PO3). Source localization results demonstrated a posterior medial source, bilateral, but predominant in the medial part of the right occipitoparietal junction, thus contralateral to scalp-EEG IIS. These noninvasive data were confirmed by the SEEG. Indeed, intracerebral IIS were posterior, medial, and bilateral, but they predominated and were associated with a slowing of background activity in the medial part of the right occipitoparietal junction (above the anterior part of the right calcarine fissure) (Table 3 and Fig. 1). Source localization for lateral group Two patients (Patients 3 and 4) were classified in this group and were characterized by subcontinuous scalpEEG IIS (Table 1). For Patient 3, source localizations were very accurate, localized in the right temporoparietooccipital junction and validated by SEEG data (Table 3 and Fig. 3). For Patient 4, source localization corresponded to the main part of the intracerebral interictal distribution, which was bilateral and predominated in right lateral posterior structures (Tables 2 and 3). Source localization for mediolateral group Seven patients were classified in this group (Patients 5–11). Organization of intracerebral IIS was characterized by its complexity, being bilaterally distributed for 4 of 7 patients (Tables 2 and 3), with an implication of both medial and lateral posterior structures, and an implication of medial temporal lobe structures for 5 of 7 patients (medial contacts of depth electrodes A, B, C for Patients 5, 6, 9, 10, 11; Table 2). The medial posterior contingent was accurately localized for 4 of 7 patients (Patients 5, 6, 9, 10; Table 3). For Patient 5, all source localizations were located in the right occipital lobe (medial and lateral parts). These source localizations corresponded only to a part of the intracerebral interictal distribution, which was quite widespread for this patient. Intracerebral IIS were indeed posterior bilateral and for each occipital lobe, intracerebral IIS were in both medial and lateral structures (Table 2). Moreover, there was bilateral medial temporal intracerebral IIS. Left occipital and medial temporal intracerebral IIS were nonlocalized. Patient 6 was characterized by two types of scalp-EEG IIS. For the first type, source localizations were bilateral medial posterior. For the second type, source localizations Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

284 M. Gavaret et al.

Figure 2. Representative example of each intracerebral interictal distribution type in PCE. Intracerebral data are presented with a bipolar montage. Medial posterior intracerebral IIS: In this example (Patient 2), intracerebral IIS were bilateral and strictly medial, being subcontinuously recorded from medial contacts of posterior electrodes GC (right anterior calcarine fissure), PI (ventral part of the right parietal lobe), PS (dorsal part of the right parietal lobe), GC¢ (isthmus of the left cingulate gyrus at the level of the anterior calcarine fissure), and PS’ (dorsal part of the left parietal lobe). A constant slowing of the background activity was recorded from medial contacts of GC. Lateral posterior intracerebral IIS: In this example (Patient 3), intracerebral IIS were strictly lateral, being subcontinuously recorded from lateral contacts of all depth electrodes of the right temporoparietooccipital junction (CuA, CuP, GL, OT, and GC). Intracerebral IIS were predominant and associated with a slowing of the background activity on lateral contacts of the depth electrode CuA (dorsolateral areas of the right occipital lobe) and OT (occipitotemporal junction). Mediolateral posterior intracerebral IIS: In this example (Patient 8), intracerebral IIS were medial (medial contacts of depth electrodes called CuA and GC) and independently lateral (lateral contacts of depth electrodes called GC, CuP, and GL). Epilepsia ILAE

indicated the right temporoparietooccipital junction. These source localizations were adequate but were only a part of a very large intracerebral interictal distribution, which was bilateral and mediolateral for the right occipital lobe. Bilateral medial temporal lobe intracerebral IIS were nonlocalized. Patient7wasalsocharacterizedbytwotypesofscalp-EEG IIS. For the first type, source localizations indicated the Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

lateral part of the right temporal lobe. For the second type, source localizations were in medial and basal parts of the left occipital lobe. These data were not confirmed by the SEEG, which demonstrated intracerebral IIS predominant in lateral and medial structures of the right occipital lobe and in lateral cortexoftheleftoccipitallobe. Patient 8 was characterized by usual absence of scalpEEG IIS. No source localization was therefore obtained.

10 2 types

— 10

10

10

7-BL

8-FC 9-ZC

10-ML

11-VC

R temporal pole/R TL

R lateral TL L medial and basal OL — L (medial/lateral/basal) OL L temporal posterior cortex R cuneus

Bilateral medial posterior R TPO junction

0.85–0.92

0.92–0.96

— 0.83–0.93

0.95–0.97

0.88–0.95

R TL

R lateral TL L medial and basal OL — L (medial/lateral/basal) OL and L temporal posterior cortex R cuneus

Bilateral medial posterior R TPO junction

Bilateral OL (mediolateral) Bilateral TL (medial) R mediolateral OL, medial PL, and medial TL L medial OL, PL, TL Bilateral lateral OL R medial OL R mediolateral OL and PL Lateral: L, OL, and posterior TL Medial: L, TL, L, OL R medial and lateral OL R medial TL Bilateral mediolateral OL and TL

R lateral TPO junction Lateral, bilateral, predominant in R TPO junction R medial TL

Medial left cuneus Bilateral medial posterior with a right predominance

Summarized intracerebral IIS distribution

IIS source localization results and summarized intracerebral interictal distribution for each patient. IIS, interictal spikes; L, left; MUSIC, multiple signal classification; OL, occipital lobe; PL, parietal lobe; R, right; TL, temporal lobe; TPO junction, temporoparietooccipital junction.

10 2 types

6-RC

0.89–0.97

Mediolateral posterior intracerebral IIS (posterior M/L index: 0.5–4) 5-PS 10 R OL (mediolateral) R OL (mediolateral)

R lateral TPO junction R lateral occipitotemporal junction

0.86–0.94 0.89–0.96

MUSIC

— Bilateral medial posterior with a right predominance

Overall correlation

Equivalent current dipole (ECD) Localization

— 0.86–0.96

Scalp EEG IIS

Medial posterior intracerebral IIS (posterior M/L index > 4) 1-LF — — 2-LN 10 Bilateral medial posterior with a right predominance Lateral posterior intracerebral IIS (posterior M/L index < 0.5) 3-BA 10 R lateral TPO junction 4-DS 10 R lateral occipitotemporal junction

Patients

Table 3. IIS source localization results

285

Source Localization in Posterior Epilepsy

Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

286 M. Gavaret et al.

Figure 3. Scalp EEG, source localization and intracerebral IIS for Patient 3 (lateral posterior group). (A) Scalp-EEG IIS were subcontinuously recorded over both posterior regions, their amplitude being maximal on channels TP8, PO10, P8 (monopolar montage, common average reference). For source localization, in order to avoid as possible propagated interictal activities, we have selected first interictal event, at the beginning of a rhythmic burst. (B) Amplitude cartography (using Focus software) during the single IIS surrounded in panel A. (C) The same single IIS, 64 channels superimposed. SNR was superior to 2. Temporal window of analysis lies between the two vertical lines. (D) Interictal source localization with a rotating dipole (red arrow) was localized in the basal part of the right temporoparietooccipital junction. Interictal source localization with MUSIC indicated the lateral and basal part of the right temporoparietooccipital junction. (E) SEEG schema, superimposed upon patient’s 3D MRI, was performed using 8-depth electrodes in the right hemisphere, recording as follows (medial/lateral contacts): CuP (cuneus/angular gyrus), CuA (anterior calcarine fissure/dorsolateral occipital lobe), GL (lingual gyrus/lateral occipitotemporal junction), OT (occipitotemporal junction), GC (isthmus of the cingulate gyrus/supramarginalis gyrus), B (hippocampal head/middle temporal gyrus), C (hippocampal tail/middle temporal gyrus), TB (temporobasal areas). Intracerebral IIS were strictly lateral, involving all lateral contacts of the depth electrodes localized in the right temporoparietooccipital junction (CuA, CuP, GL, and OT). Epilepsia ILAE

SEEG implantation was unilateral. Intracerebral IIS were mediolateral in the occipital lobe and in the posterior part of the parietal lobe. For Patient 9, source localizations were in medial, basal, and lateral parts of the left occipital lobe, overlapping on left lateral posterior temporal lobe structures. These results were concordant with intracerebral data. Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

However, SEEG demonstrated left medial temporal intracerebral IIS, which were nonlocalized. For Patient 9, we did not identify several interictal spikes subtypes. As for all patients, 10 single interictal spikes were studied with source localization tools. Either with MUSIC or with equivalent current dipoles, results were localized in medial, basal, and lateral parts of the left occipital lobe,

287 Source Localization in Posterior Epilepsy overlapping with left lateral posterior temporal lobe structures. These results were concordant with intracerebral data. However, SEEG demonstrated left medial temporal intracerebral IIS that were nonlocalized. For Patient 10, source localizations (in the right cuneus) were validated, but were only a part of the intracerebral interictal distribution that was medial and lateral in the right occipital lobe. Medial temporal intracerebral IIS were nonlocalized. For Patient 11, source localizations had designated only the temporal contingent of the intracerebral interictal distribution. In summary, for these seven patients included in the mediolateral group, source localizations indicated only a part, medial or lateral, of the intracerebral interictal distribution (Table 3) underestimating its complexity and spread. Medial temporal intracerebral IIS were frequent (5 of 7 patients) and never localized.

Discussion The main objective of this study was to determine in which conditions intracerebral IIS were localized with the most accuracy and precision in posterior cortex epilepsies (PCE) cases. Results indicate that source localization accuracy depends on the intracerebral interictal distribution. All PCE patients studied here were investigated by SEEG. For 8 of 11 patients, SEEG implantation was bilateral. For six of these eight patients, SEEG demonstrated that intracerebral IIS were effectively bilateral, creating particularly difficult conditions for source localization tools. For these six patients with a bilateral posterior interictal distribution, source localization results were precise and exhaustive for only one patient, characterized by bilateral but very focal intracerebral IIS above the anterior part of the calcarine fissure (Patient 2; Fig. 1). Beside the problem of the bilateral extension of intracerebral IIS, SEEG data have clearly shown that IIS have a more complex organization than a simple focus of spikes. In the majority of PCE patients, IIS distribution corresponded to a mediolateral distribution. This ‘‘network’’ organization of spikes in posterior cortex epilepsies is reminding of what we previously described in temporal lobe epilepsies (Bourien et al., 2005). Source localizations in these cases designated only a part of the intracerebral IIS, either medial or lateral, underestimating complexity and spread of the intracerebral interictal distribution. In the literature, certain electric source imaging studies included some PCE cases (Boon et al., 1997; Huppertz et al., 2001; Merlet & Gotman, 2001; Lantz et al., 2003; Michel et al., 2004a; Sperli et al., 2006; Leal et al., 2007). Two of these studies were validated by intracranial recordings (Boon et al., 1997; Merlet et Gotman, 2001), but the number of PCE cases was too small for general conclusions. Some studies described a level of accuracy down to a sublobar level, but results were indirectly validated by

comparison with other imaging techniques or seizure outcomes (Huppertz et al., 2001; Lantz et al., 2003; Michel et al., 2004a; Sperli et al., 2006). In Michel et al. (2004a), one of two parietooccipital epilepsy patients had a lesion close to the midline. Source localization results were posterior, paramedian, and contralateral to the supposed epileptogenic lesion. One recent study (Leal et al., 2007) showed more precise results, demonstrating with LORETA epileptic activity originating near the lateral occipital area and spreading to cortical temporal or parietal areas. These results were not validated by intracerebral data, being obtained in idiopathic occipital epilepsies. In our study, only two PCE patients were characterized by predominantly medial posterior intracerebral IIS (medial group). Scalp-EEG IIS were characterized by higher amplitude contralateral to intracerebral IIS for one of these patients. This medial source, although contralateral to scalp-EEG IIS, was accurately localized by HR-EEG. This result need to be confirmed on a larger number of cases. It is however in agreement with previous studies. Indeed, it has been showed that medial occipital lobe epilepsies appeared to be characterized by contralateral scalp-EEG IIS more frequently than lateral occipital lobe epilepsies (Blume et al., 2005, in a study of 31 occipital lobe epilepsy patients investigated by scalp-EEG and grids). Moreover, it has been previously demonstrated that ‘‘the higher voltage discharge may appear on the side opposite the lesion itself’’ and noticed that this EEG finding is in particular observed ‘‘in patients with epileptogenic foci on the mesial surface of one hemisphere’’ (Tukel & Jasper, 1952). Some important pitfalls of dipole source localization arise from the procedure of fitting the simplistic singledipole model to real cortical sources with spatial extent and complex configuration (Kobayashi et al., 2005). We therefore used two different localization algorithms: unconstrained equivalent current dipole as described by Scherg (1992) and MUSIC, described by Mosher et al. (1992). The latter algorithm constructs a metric that gives a distributed view of the contribution of each generator. MUSIC produces results more likely to reflect the clinical picture, being distributed in space. We did not obtain discrepancies between these two methods. Concerning propagation of interictal activities, by studying a single signal component, with a temporal window of analysis corresponding to the IIS duration (60–110 ms), we never demonstrated propagation with a single equivalent current dipole type moving dipole, as validated by intracerebral data. Generally, at the beginning of the temporal window (HR-EEG data of low SNR), the moving dipole has an unstable position (and a low correlation). In the middle of the temporal window, at the time corresponding to the surface IIS maximal amplitude (maximal SNR), the moving dipole has a stable position (and a maximal correlation). At the end of the temporal window (when EEG signals join the isoelectric line, HR-EEG data of low SNR), the Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

288 M. Gavaret et al. moving dipole has again an unstable position (and a low correlation). To improve source localization performance, this would require further methodological study comparing different inverse problem algorithms, notably distributed and dipolar source models and comprising distinct temporal window analysis (rising phase of interictal activity versus whole interictal activity duration). In conclusion, the majority of PCE patients were characterized by widespread and complex intracerebral IIS, being most frequently distributed bilaterally and mediolaterally. In these cases, source localization tools used in this study were insufficient to describe the complexity of the interictal distribution. These findings show that the organization of the ‘‘irritative’’ zone, perhaps particularly in posterior cortex epilepsies, cannot be naively considered as a simple focus restricted to a lesional site. However, we showed that for PCE, whose intracerebral IIS had a less spatial extent, whether lateral or medial, source localization results were accurate.

Acknowledgments Conflict of interest: We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. The authors have no conflicts of interest to disclose.

References Bancaud J. (1969) Les crises pileptiques d’origine occipitale (tude stro-lectroencphalographique) Rev Oto-neuro-ophtalmol 41: 299–314. Barba C, Doglietto F, De Luca L, Faraca G, Marra C, Meglio M, Rossi GF, Colicchio G. (2005) Retrospective analysis of variables favouring good surgical outcome in posterior epilepsies. J Neurol 252:465– 472. Basseville M, Nikiforov IV. (1993) Detection of abrupt changes: theory and application. Prentice-Hall, Englewood Cliffs, NJ. Bautista RE, Cobbs MA, Spencer DD, Spencer SS. (1999) Prediction of surgical outcome by interictal epileptiform abnormalities during intracranial EEG monitoring in patients with extrahippocampal seizures. Epilepsia 40:880–890. Blume WT, Whiting SE, Girvin JP. (1991) Epilepsy surgery in the posterior cortex. Ann Neurol 29:638–645. Blume WT, Wiebe S, Tapsell LM. (2005) Occipital epilepsy: lateral versus mesial. Brain 128:1209–1225. Boesebeck F, Schulz R, May T, Ebner A. (2002) Lateralizing semiology predicts the seizure outcome after epilepsy surgery in the posterior cortex. Brain 125(Pt 10):2320–2331. Boon P, D’Have M, Adam C, Vonck K, Baulac M, Vandekerckhove T, De Reuck J. (1997) Dipole modeling in epilepsy surgery candidates. Epilepsia 38:208–218. Bourien J, Bellanger JJ, Bartolomei F, Chauvel P, Wendling F. (2004) Mining reproducible activation patterns in epileptic intracerebral EEG signals: application to interictal activity. IEEE Trans Biomed Eng 51:304–315. Bourien J, Bartolomei F, Bellanger JJ, Gavaret M, Chauvel P, Wendling F. (2005) A method to identify reproducible subsets of co-activated structures during interictal spikes. Application to intracerebral EEG in temporal lobe epilepsy. Clin Neurophysiol 116:443–455. Braak H. (1977) The pigment architecture of the human occipital lobe. Anat Embryol 150:229–250. Chatrian GE, Bergamini L, Dondey M, Klass DW, Lennox-Buctal M, Petersen I. (1974) A glossary of terms most commonly used by Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x

clinical electroencephalographers. Electroencephalogr Clin Neurophysiol 37:538–548. Clarke S, Miklossy J. (1990) Occipital cortex in man: organization of callosal connections, related myelo- and cytoarchitecture, and putative boundaries of functional visual areas. J Comp Neurol 298:188– 214. Clarke S. (1994) Modular organization of human extrastriate visual cortex: evidence from cytochrome oxidase pattern in normal and macular degeneration cases. Eur J Neurosci 6:725–736. Dalmagro CL, Bianchin MM, Velasco TR, Alexandre V, Walz R, Terra-Bustamante VC, Inuzuka LM, Wichert-Ana L, Araujo D, Serafini LN, Carlotti CG, Assirati JA, Machado HR, Santos AC, Sakamoto AC. (2005) Clinical features of patients with posterior cortex epilepsies and predictors of surgical outcome. Epilepsia 46: 1442–1449. Daniel RT, Meagher-Villemure K, Farmer JP, Andermann F, Villemure JG. (2007) Posterior quadrantic epilepsy surgery: technical variants, surgical anatomy, and case series. Epilepsia 48:1429–1437. Gavaret M, Badier JM, Marquis P, Bartolomei F, Chauvel P. (2004) Electric source imaging in temporal lobe epilepsy. J Clin Neurophysiol 21:267–282. Gavaret M, Badier JM, Marquis P, McGonigal A, Bartolomei F, Regis J, Chauvel P. (2006) Electric source imaging in frontal lobe epilepsy. J Clin Neurophysiol 23:358–370. Geddes LA, Baker LE. (1967) The specific resistance of biological materials—a compendium of data for the biomedical engineer and physiologist. Med Biol Eng 5:271–293. GonÅalves S, de Munck JC, Heethaar RM, Lopes da Silva FH, van Dijk BW. (2000) The application of electrical impedance tomography to reduce systematic errors in the EEG inverse problem—a simulation study. Physiol Meas 21:379–393. Huppertz HJ, Hof E, Klisch J, Wagner M, Lucking CH, Kristeva-Feige R. (2001) Localization of interictal delta and epileptiform EEG activity associated with focal epileptogenic brain lesions. Neuroimage 13:15–28. Kobayashi K, Yoshinaga H, Ohtsuka Y, Gotman J. (2005) Dipole modelling of epileptic spikes can be accurate or misleading. Epilepsia 46:397–408. Kuzniecky RI, Barkovich AJ. (2001) Malformations of cortical development and epilepsy. Brain Dev 23:2–11. Lantz G, Spinelli L, Seeck M, Grave de Peralta R, Sottas C, Michel C. (2003) Propagation of interictal epileptiform activity can lead to erroneous source localizations: a 128-channel EEG mapping study. J Clin Neurophysiol 20:311–319. Leal AJ, Nunes S, Dias AI, Vieira JP, Moreira A, Calado E. (2007) Analysis of the generators of epileptic activity in early-onset childhood benign occipital lobe epilepsy. Clin Neurophysiol 118:1341– 1347. Lee SK, Lee SY, Kim DW, Lee DS, Chung CK. (2005) Occipital lobe epilepsy: clinical characteristics, surgical outcome, and role of diagnostic modalities. Epilepsia 46:688–695. Ludwig BI, Ajmone Marsan C. (1975) Clinical ictal patterns in epileptic patients with occipital electroencephalographic foci. Neurology 25:463–471. McGonigal A, Bartolomei F, Regis J, Guye M, Gavaret M, Trebuchon-Da Fonseca A, Dufour H, Figarella-Branger D, Girard N, Peragut JC, Chauvel P. (2007) Stereoelectroencephalography in presurgical assessment of MRI-negative epilepsy. Brain 130:3169–3183. Merlet I, Gotman J. (2001) Dipole modelling of scalp electroencephalogram epileptic discharges: correlation with intracerebral fields. Clin Neurophysiol 112:414–430. Michel C, Lantz G, Spinelli L, Grave de Peralta R, Landis T, Seeck M. (2004a) 128-channel EEG source imaging in epilepsy: clinical yield and localization precision. J Clin Neurophysiol 21:71–83. Michel C, Murray M, Lantz G, Gonzalez S, Spinelli L, Grave de Peralta R. (2004b) EEG source imaging. Clin Neurophysiol 115:2195– 2222. Mosher JC, Lewis PS, Leahy RM. (1992) Multiple dipole modeling and localization from spatio-temporal MEG data. IEEE Trans Biomed Eng 39:541–557. Mosher JC, Baillet S, Leahy RM. (1999) EEG source localization and imaging using multiple signal classification approaches. J Clin Neurophysiol 16:225–238.

289 Source Localization in Posterior Epilepsy Olivier A, Gloor P, Andermann F, Ives J. (1982) Occipitotemporal epilepsy studied with stereotaxically implanted depth electrodes and successfully treated by temporal resection. Ann Neurol 11:428–432. Ono M, Kubik S, Abernathey CD. (1990) Atlas of the cerebral sulci. Thieme Medical Publishers, New York. Palmini A, Andermann F, Dubeau F, Gloor P, Olivier A, Quesney LF, Salanova V. (1993) Occipitotemporal epilepsies: evaluation of selected patients requiring depth electrodes studies and rationale for surgical approaches. Epilepsia 34:84–96. Palmini A, Andermann F, Dubeau F, da Costa JC, Calcagnotto ME, Gloor P, Olivier A, Paglioli E, Paglioli-Neto E. (1999) Occipitotemporal relations: evidence for secondary epileptogenesis. Adv Neurol 81:115–129. Salanova V, Andermann F, Olivier A, Rasmussen T, Quesney LF. (1992) Occipital lobe epilepsy: electroclinical manifestations, electrocorticography, cortical stimulation and outcome in 42 patients treated between 1930 and 1991. Brain 115:1655–1680. Scherg M. (1992) Functional imaging and localization of electromagnetic brain activity. Brain Topogr 5:103–111. Senhadji L, Dillenseger JL, Wendling F, Rocha C, Kinie A. (1995) Wavelet analysis of EEG for three-dimensional mapping of epileptic events. Ann Biomed Eng 23:543–552. Sperli F, Spinelli L, Seeck M, Kurian M, Michel C, Lantz G. (2006) EEG source imaging in pediatric epilepsy surgery: a new perspective in presurgical workup. Epilepsia 47:981–990.

Sveinbjornsdottir S, Duncan JS. (1993) Parietal and occipital lobe epilepsy: a review. Epilepsia 34:493–521. Talairach J, Bancaud J, Szikla G, Bonis A, Geier S, Vedrenne C. (1974) New approach to the neurosurgery of epilepsy. Stereotaxic methodology and therapeutic results. Neurochirurgie 20:1–240. Tao JX, Baldwin M, Hawes-Ebersole S, Ebersole JS. (2007) Cortical substrates of scalp EEG epileptiform discharges. J Clin Neurophysiol 24:96–100. Tukel K, Jasper H. (1952) The electroencephalogram in parasagittal lesions. Electroencephalogr Clin Neurophysiol 4:481–494. Westmoreland BF. (1998) The EEG findings in extratemporal seizures. Epilepsia 39(Suppl 4):S1–S8. Williamson PD, Thadani VM, Darcey TM, Spencer DD, Spencer SS, Mattson RH. (1992) Occipital lobe epilepsy: clinical characteristics, seizure spread patterns, and results of surgery. Ann Neurol 31:3–13. Zanow F, Peters MJ. (1995) Individually shaped volume conductor models of the head in EEG source localisation. Med Biol Eng Comput 33:582–588. Zumsteg D, Friedman A, Wennberg RA, Wieser HG. (2005) Source localization of mesial temporal interictal epileptiform discharges: correlation with intracranial foramen ovale electrode recordings. Clin Neurophysiol 116:2810–2818.

Epilepsia, 50(2):276–289, 2009 doi: 10.1111/j.1528-1167.2008.01742.x