What is the significance of interictal water diffusion changes in frontal lobe epilepsies?

www.elsevier.com/locate/ynimg NeuroImage 35 (2007) 28 – 37

What is the significance of interictal water diffusion changes in frontal lobe epilepsies? M. Guye, a,b,c,d,⁎ J.P. Ranjeva, b,c F. Bartolomei, a,c,d S. Confort-Gouny, b,c A. McGonigal, a,c,d J. Régis, a,c,e P. Chauvel, a,c,d and P.J. Cozzone b,c a

INSERM, U 751, Laboratoire de Neurophysiologie et Neuropsychologie, Marseille, France CNRS, UMR 6612, Centre de Résonance Magnétique Biologique et Médicale (CRMBM), Marseille, France c Université de la Méditerranée, Faculté de Médecine, Marseille, France d CHU Timone, Service de Neurophysiologie Clinique, Marseille, France e CHU Timone, Service de Neurochirurgie Fonctionnelle, Marseille, France b

Received 12 September 2006; revised 1 November 2006; accepted 13 November 2006 Available online 18 January 2007 The aim of this study was to better understand the significance of interictal changes in water molecule diffusivity defined by diffusionweighted imaging (DWI) in frontal lobe epilepsy (FLE), as well as to test the accuracy of interictal DWI in the definition of the epileptogenic zone (EZ). DWI was carried out in 14 patients with refractory FLE (9 negative-MRI) as well as in 25 controls. Statistical mapping analysis (SPM2) of diffusivity maps was used to detect, for each subject, significant diffusivity alterations. We then studied the relationships between diffusion and depth recorded electrical abnormalities. Clinical correlates of the extent of diffusivity changes were also tested. We found areas of significantly increased diffusivity (SID) in 13 patients. Eight had SID in the EZ, 9 within the irritative zone (IZ) and 12 outside, mainly in connected areas. We found a correlation between the extent of SID and the duration of epilepsy (p corrected = 0.026, R = 0.621). In addition, SID was significantly less widespread in negative-MRI patients (p = 0.028). However, we found no significant differences concerning either seizure frequency (p = 0.302), seizure generalization (p = 0.841), history of status (p = 0.396), or surgical outcome (p = 0.606). We suggest that SID in normal appearing areas is not a specific signature of epileptogenicity in FLE, and is more likely to reflect multifactorial and potentially evolving neuro-glial injuries. © 2006 Elsevier Inc. All rights reserved. Keywords: Frontal lobe epilepsy; Diffusion imaging; Statistical mapping analysis; Stereo-electro-encephalography

⁎ Corresponding author. Laboratoire de Neurophysiologie et Neuropsychologie, INSERM U 751 and Service de Neurophysiologie Clinique, Faculté de Médecine et CHU Timone, 27 boulevard Jean Moulin, 13005 Marseille, France. Fax: +33 4 91 25 65 39. E-mail address: [email protected] (M. Guye). Available online on ScienceDirect (www.sciencedirect.com). 1053-8119/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2006.11.049

Introduction Frontal lobe epilepsy (FLE) is the second most common form of drug-resistant partial epilepsy referred to epilepsy surgery centres (Rosenow and Luders, 2001). However, though frequent, intractable FLEs are still poorly studied and remain ill-understood. Indeed, the epileptogenic zone (EZ) that must be surgically removed in order to cure patients is much more difficult to define in FLE than in temporal lobe epilepsy (TLE) (Chauvel et al., 1992, 1995; Jobst and Williamson, 2005; Williamson and Jobst, 2000). This may be explained by several anatomical and functional differences between FLE and TLE: (i) the size of the frontal lobe, (ii) the complexity and wide connectivity of the neural systems potentially involved in frontal lobe seizures, and (iii) the often misleading electroclinical patterns due to fast discharge spread. Therefore, intracranial recordings are frequently required in the presurgical evaluation of FLEs (Talairach et al., 1992). However, invasive investigations are still difficult to perform in such epilepsies and need to be guided. In this context, better definition of structural alterations and their epileptogenicity in vivo, are major issues. Structural changes at a cellular scale may be assessed by diffusion-weighted imaging (DWI), which is a non-invasive MR technique allowing quantification of passive water motion or diffusivity (expressed as apparent diffusion coefficient (ADC) or mean diffusivity (MD)). In the brain, this diffusivity, being mainly extra-cellular, is constrained by cell membranes and organelles, axons and myelin sheaths (Basser and Pierpaoli, 1996). Therefore, DWI measures the interactions between water molecules and cerebral structures. Thus, any cell or tissue alteration may affect water molecule motion and consequently be assessed in vivo by measuring water diffusivity. Using DWI, animal studies have first demonstrated an early and transient diffusivity decrease during provoked status or sustained seizures (Ebisu et al., 1996; Nakasu et al., 1995; Prichard et al.,

M. Guye et al. / NeuroImage 35 (2007) 28–37

1995; Zhong et al., 1993, 1997). This decrease has been interpreted as a functional cellular swelling due to the edema produced by seizure excitotoxicity (Wang et al., 1996). In human studies of status epilepticus, DWI data have shown more complex modifications comprising both decreased and increased diffusivity (Hisano et al., 2000; Wieshmann et al., 1997). Peri-ictal and post-ictal human studies using DWI or diffusion tensor imaging (DTI) (i.e. a diffusion imaging technique allowing quantification of water diffusivity as well as anisotropy) showed in some cases, transiently decreased local diffusivity, potentially in concordance with the EZ (Diehl et al., 1999, 2001, 2005; Hufnagel et al., 2003). Conversely, interictal diffusion imaging is more likely to demonstrate fixed tissue changes susceptible to affect water diffusivity. Several diffusion studies have already demonstrated areas of significantly increased diffusivity (SID) in a proportion of patients with partial epilepsies during the interictal period (Arfanakis et al., 2002; Assaf et al., 2003; Eriksson et al., 2001; Rugg-Gunn et al., 2001, 2002; Sundgren et al., 2004; Thivard et al., 2005, 2006; Lee et al., 2004; Kantarci et al., 2002). Numerous studies have focused on TLE demonstrating SID in mesial temporal lobe structures, predominating on the ipsilateral side in individual and group studies (Assaf et al., 2003; Lee et al., 2004; Kantarci et al., 2002; Arfanakis et al., 2002; Thivard et al., 2005). However, individual approaches have also shown SID in areas outside the temporal lobe (Arfanakis et al., 2002; Thivard et al., 2005). Rugg-Gunn et al. were the first to demonstrate areas of SID in 8 out of 30 negative-MRI patients presenting with several forms of partial epilepsy. Six SID were concordant with surface EEG abnormalities (Rugg-Gunn et al., 2001). In that study, the clinical data provided (no depth recording) showed that 8 of the 30 patients probably suffered from FLE. Three out of these 8 probable FLEs had SID compatible with the surface EEG abnormalities. In a case report, the same group reported good concordance between a single area of SID and the EZ identified by using depth electrode recording in one patient (Rugg-Gunn et al., 2002). Thivard and colleagues found areas of SID in 11 out of 16 patients, including 2 cases of FLE (Thivard et al., 2006). Performing a careful comparison between diffusivity and electrical abnormalities by using depth electrode recordings, they found congruent location in 7 out of the 11 patients (including 1 FLE). Therefore, only 3 cases of careful comparison between diffusion imaging and depth electrode recording in FLE are currently available in the literature. Although comparisons between the location of diffusivity abnormalities and epileptogenic areas have been already addressed in partial epilepsy, the significance of such abnormalities has not been fully examined in FLE. In the present study, we aimed to better understand the relationships between interictal water diffusivity changes and electrical abnormalities in FLE. We also aimed to assess the accuracy of interictal DWI in the definition of the EZ in such epilepsies. We studied 14 patients suffering from intractable FLE (9 MRI-negative). DWI of each patient was compared to a group of 25 control subjects on a voxel-by-voxel basis to define statistically abnormal areas of diffusivity. We then compared the location of diffusivity changes with the location of depth recorded electrical abnormalities (defined by using stereo-electroencephalography (SEEG)). Based on these comparisons, we studied the specificity and sensitivity of DWI in the definition of the epileptogenic zone. We also tested whether the extent of diffusion abnormalities was correlated to clinical features.

29

Materials and methods Subjects We studied 14 patients with drug-resistant frontal lobe epilepsy (10 women, 4 men; mean age 28 years, range from 14 to 41), who were candidates for epilepsy surgery and for whom invasive recording was indicated. Table 1 summarizes the clinical features of the 14 patients as well as the post-surgical pathology and outcomes when available. The control group was formed by 25 healthy controls with no history of neurological disorder (15 women, 10 men; mean age 27 years, range from 18 to 50) in order to obtain normative diffusion data. All patients and healthy volunteers gave their informed consent to this protocol, which was approved by the Ethics Committee of the Timone Hospital in Marseille, France. MRI acquisition MRI was performed in all patients and healthy volunteers, using a Magnetom Vision Plus system (1.5 T) (Siemens Erlangen Germany). The conventional MRI protocol included localizer scout images and the following sequences: – T1-weighted images (TE/TR = 15/700 ms, 23 slices, 5 mm slice thickness, field of view (FOV) = 240 mm, matrix 256 × 256) acquired in the AC–PC plane, – T2-weighted images (TE/TR = 112/7308 ms, FOV = 240 mm, matrix 256 × 256, 23 slices, 5 mm slice thickness) acquired in the AC–PC plane, – T1-weighted inversion recovery images (TE/TR = 60/8000 ms, TI = 350 ms, FOV = 240 mm, matrix 512 × 512, 5 mm slice thickness), in a coronal plane perpendicular to the AC–PC plane, – FLAIR images (TE/TR = 110/8000 ms, TI = 2500 ms, FOV = 240 mm, matrix 256 × 256, 5 mm slice thickness) in a coronal plane perpendicular to the AC–PC plane, – 3D-MPRAGE images (TE/TR = 4/9.7 ms, isotropic voxel of 1.25 mm3). Diffusion imaging Diffusivity of water molecules was assessed by calculating mean diffusivity (MD) maps (average ADC values in each direction), obtained in the axial plane from a diffusion-weighted echo planar imaging (EPI) sequence (TE = 100 ms, 19 slices, thickness = 5 mm, FOV = 240 mm, matrix = 1282, b = 0; 250; 500; 1000 s/mm2 acquired sequentially in the x, y and z directions). Diffusion imaging was acquired at least 6 h after the last seizure even in patients with high seizure frequency. Diffusivity image processing EPI images (b = 0) were spatially normalized into the MNI space using the EPI anatomical template provided by SPM2 and transformations were applied onto the corresponding MD maps that were re-sampled using a 2 × 2 × 2 mm3 kernel pixel size (http:// www.fil.ion.ucl.ac.uk/spm/software). Spatial normalization algorithm preserved voxel intensities (concentrations) whatever region volumes were stretched by warping (Ashburner and Friston, 2000; Good et al., 2001). Before statistical comparison, MD maps were

30

M. Guye et al. / NeuroImage 35 (2007) 28–37

Table 1 Clinical features, conventional MRI, pathology and surgical outcomes when available Patients Sex Age Medical history

Epilepsy Seizure duration (y) frequency

Secondary Conventional generalization MRI

Pathology

Surgical outcome

Taylor type FCD + microdysgenesis Taylor type FCD + gliosis Taylor type FCD + gliosis Taylor type FCD ∅ ∅ Gliosis ∅ ∅ Taylor type FCD Non-specific + gliosis

Engel IA Engel IA Engel IIIB Engel IA ∅ ∅ Engel IA ∅ ∅ Engel IC Engel IA

∅

∅

Gliosis

Engel IA

Taylor type FCD + macrodysgenesis

Engel IIA

1

F

27

–

11

7/w

–

Negative

2 3 4 5 6 7 8 9 10 11

F F M M F F F M F F

19 14 33 26 20 41 18 29 25 31

– – – – – – – FC at 9 months 2 FC at 4 y –

14 10 17 16 17 21 8 17 20 21

10–100/w 10–20/w 3 status 7–28/w 7/w 1–30/w 2–3 status/y 10–15/w 6–7/w 2–3 status/y 7/w 1–10/w 1 status 10/w

<5 – 1–2/y 2/month – – – – <1/y –

12

F

41

23

1–2/w

–

13

M

31

19

1–2/w

<5

14

F

29

Fronto-orbit lesion at 18 y Fronto-orbit trauma at 12 y –

25

2/w

1–5/y

Negative Negative Negative Negative Negative Negative Negative Negative L frontal MCD R post-central MCD R orbit post-surgical lesion L orbit post-traumatic lesion L front dorso-lat MCD

FC: febrile convulsion; y: year; w: week; R: right, L: left; orbit.: orbital; dorso-lat: dorsolateral; MCD: malformation of cortical development; FCD: focal cortical dysplasia; –: no; ∅: surgery not performed. The surgical outcomes are presented following Engel's classification: Class I, free of disabling seizure (IA, completely seizure free since surgery; IC, some disabling seizures after surgery but free of disabling seizure since at least 2 years); Class II, rare disabling seizures (IIA, rare disabling seizures since surgery); Class III, worthwhile improvement (IIIB: prolonged seizure-free intervals amounting to greater than half the followed-up period, but not<2 years); Class IV: no worthwhile improvement.

smoothed using a 8 mm Gaussian filter (about four times the spatial resolution) to minimize remaining spatial difference between subjects and to better satisfy conditions of the random field theory (Friston et al., 1995). Statistical mapping of diffusivity data We used SPM2 software for statistical comparison. An Ancova, with age as confounding covariate (p < 0.001, k = 5; FDR corrected p < 0.05 at a voxel level meaning that, on average, 5% of the voxels reported as showing a significant effect will be false-positive), was performed on a voxel-by-voxel basis to compare diffusivity maps of each patient with diffusivity maps of the control group. Two contrasts were used to determine whether each voxel of each subject had increased or decreased diffusivity in comparison with the control group. We first compared each control with the remaining control group to evaluate the individual variability in controls. We then compared each patient with the control group. Clusters were located on an atlas after transformation of MNI coordinates into Talairach coordinates (mni2tal toolbox for the SPM2 software). Presurgical evaluation—SEEG recording All patients underwent non-invasive investigations (referred to as phase I) including video-EEG recordings of seizures (electroclinical correlations), neuropsychological evaluation, interictal spike localization using high-resolution EEG (i.e. 64 electrodes), interictal ± ictal single-photon emission computed tomography (SPECT) and structural MRI. All then underwent a phase II evaluation consisting of stereoelectroencephalography (SEEG) (Munari et al., 1995; Chauvel

et al., 1996; Talairach et al., 1992; Guenot et al., 2001; Bartolomei et al., 2002, 2004, 2005; Cossu et al., 2005). This method consists of an implantation of multiple-lead intracerebral electrodes (diameter: 0.8 mm; number of contacts: 10 to 15; contact length: 2 mm; interval between contacts: 1.5 mm) using a standard system fastened to the Talairach stereotactic frame (Talairach et al., 1974, 1992). Six to fifteen electrodes are usually implanted, providing in each case more than 100 sites of recording and consequently an extended electrophysiological sampling of the brain areas of interest. The positioning of electrodes was established in each patient based upon available non-invasive information provided by the phase I cited above. SEEG is proposed when the phase I alone does not permit adequate localisation of the EZ, and where this provides enough anatomo-electro-clinical data in order to establish prerequisite hypotheses about the localization of the EZ. The implantation accuracy was per-operatively controlled by telemetric X-ray imaging. A post-operative computerized tomography (CT) scan without contrast was used to verify the absence of bleeding and the precise location of each recording lead. Following the recording period of 3–9 days, intracerebral electrodes were then removed and an MRI performed, permitting visualization of the trajectory of each electrode. Finally, CT-scan/MRI data fusion was performed to anatomically locate each contact along the electrode trajectory (Bartolomei et al., 2004). The EZ is defined as the region(s) where the ictal discharge originates. The irritative zone (IZ) is defined as the region where interictal paroxysmal events such as spikes are recorded (Guye et al., 2005). Video-SEEG recording allowed a delineation of the EZ and IZ in all patients. According to previous studies (Guye et al., 2002, 2005), we schematically classified the SEEG results

M. Guye et al. / NeuroImage 35 (2007) 28–37

according to the 3 following criteria: (i) regions involved in the EZ; (ii) regions involved in the IZ; and (iii) regions without electrical abnormalities. These three criteria were applied to regions of interest representing key structures of the main frontal lobe regions potentially involved in frontal seizures and classically associated with Brodmann areas (BA) corresponding to functional neural systems. We thus individualized the following anatomical regions associated with one or several BA: frontal polar (BA 10-9), orbitofrontal (BA 11-25), opercular (BA 44-45-47), cingulate (BA 24), pre-cingulate (BA 32), dorsolateral prefrontal (BA 9-46), lateral pre-motor (BA 6-8), medial pre-motor (supplementary motor area or SMA (BA 6) and pre-SMA (BA 8)), pre-central (BA 4) and post-central (BA 1-2-3) (the latter located in the parietal lobe but often discharging with pre-central region during central seizures). Other regions located outside the frontal lobe but closely connected to this, such as the insula, parietal and temporal lobes, were also explored when necessary to precisely define the EZ and IZ.

31

changes (i.e. represented by the number of pixels belonging to significant clusters) and the duration of epilepsy calculated in years. We also performed non-parametric Mann–Whitney’s comparisons of the extent of significant diffusivity changes between groups presented as follows: (i) the patients presenting with a relatively low seizure frequency (arbitrarily defined as an average of seizure < 10 per week) versus those presenting with a high seizure frequency (arbitrarily defined as an average of seizure ≥ 10 per week), (ii) the patients presenting with seizures secondarily generalized versus those without generalization, (iii) the patients having experienced one or several episodes of status epilepticus in their clinical history versus those who had not; (iv) the patients presenting with visible MR lesion versus MR-negative patients; and (v) the patients post-surgically classified as Engel Class I versus those classified in other groups. Results Stereoelectroencephalography

Diffusion imaging and SEEG recording comparisons This form of classification was used to simplify and organize the SEEG results. SEEG and diffusion data comparisons were based on the same anatomical referential following Talairach Atlas coordinates. Normalization of diffusion images and transformation of MNI coordinates into Talairach coordinates allowed such comparisons. Sensitivity (Se) and specificity (Sp) of diffusion imaging as a test to define the EZ was also calculated. In order to calculate these values, we considered: (i) the number of truepositive (TP) as the number of patients having SID in the EZ, (ii) the number of false-positive (FP) as the number of patients having SID outside the EZ, (iii) the number of true-negative (TN) as the number of patients having no SID outside the EZ, and (iv) the false-negative (FN) as the number of patients having no SID in the EZ. These two values were tested successively for the whole group of patients, the “negative-MRI” group, and the “positive-MRI” group following the formula: Se = TP/(TP + FN) and Sp = TN/(TN + FP). Se and Sp were also calculated for the IZ in the same way. We also investigated Se and Sp for each individual. This approach allowed correction for the sampled volumes. In order to calculate these values, we considered for each patient: (i) the number of TP to be the number of regions involved by the EZ demonstrating SID, (ii) the number of FP to be the number of regions outside the EZ demonstrating SID, (iii) the number of TN to be the number of regions outside the EZ without SID, and (iv) the number of FN to be the number of regions involved by the EZ without SID. Se and Sp were also tested successively for the three groups of patients, following the same formula as above, and were also calculated for the IZ in the same way. Clinical correlations of diffusion abnormalities In order to define clinical correlates of diffusivity abnormalities we studied the potential relation between the extent of the areas exhibiting significant diffusivity changes and the following clinical features: (i) duration of epilepsy, (ii) estimated rate of seizure frequency, (iii) presence of secondary generalized seizures, (iv) clinical history of status epilepticus, (v) presence of visible MR lesion, and (vi) surgical outcome. We therefore performed a non-parametric Spearman Rank correlation test between the extent of significant diffusivity

Results from the depth electrical exploration of patients are detailed in Table 2. Ten patients had prefrontal epilepsy (2 with fronto-temporal epilepsy, i.e. EZ located both in anterior temporal lobe and prefrontal cortex), 2 had premotor epilepsy, 1 had central epilepsy and 1 had central and premotor epilepsy. Conventional MRI Nine out of the 14 patients had negative MRI (see Table 1). Three patients presented with visible lesions on conventional MRI related to malformations of cortical development (MCD) and two patients with visible lesions related to brain trauma. All control subjects had normal appearing conventional MRI. Statistical mapping analysis of diffusivity maps in controls No significant decreased diffusivity was observed in any control relative to the remaining group of controls. Statistical mapping analysis showed SID in only 1 out of the 25 controls. In this control, three clusters of SID were identified at a threshold of p < 0.001 (FDR corrected p < 0.05). All clusters were identified within the temporal lobe: 2 clusters in mesial temporal lobe structures (right temporal white matter near hippocampus and left temporal white matter including the amygdala) and 1 in superior temporal lobe structures (BA 22). No significant diffusivity change was found within the frontal lobe in each control in comparison with the remaining controls. Statistical mapping analysis of diffusivity maps in patients No significant cerebral decreased diffusivity was observed in each patient relative to the group of controls. Conversely SID was observed in 13 out of 14 patients (93%). All significant clusters and the locations of local maxima are shown in “Supplementary-Material_Table1” available in the online version. Diffusion and electrical abnormalities comparisons The comparisons between the location of diffusion changes and electrical abnormalities are shown in Table 3. A SID was found in

32

M. Guye et al. / NeuroImage 35 (2007) 28–37

Table 2 Summary of SEEG results SEEG Patients

Frontal pole

Front Op

Orb

Pre Cing

Cing

Pre FDorso Lat

Pre Mot Lat

SMA/ Pre-SMA

Pre Cent

Post Cent

Ins

TP

Hip

A

T

Par

Side

Contr

1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 – 2 0 2 0 0 0 2 0 1 0 1

0 0 0 2 2 2 2 – 1 2 0 2 0 0

0 0 0 0 1 1 1 0 2 0 0 2 1 2

1 0 0 0 1 0 0 0 0 0 0 0 1 0

1 0 0 0 0 0 0 0 0 0 0 2 2 2

0 0 0 1 0 2 2 0 0 2 0 1 0 2

0 1 1 0 0 2 0 1 0 1 0 0 0 2

0 2 1 0 0 0 0 1 0 1 0 0 2 2

– 0 0 – – 0 – 1 0 0 2 – 0 –

– 0 0 – – 0 – 0 0 0 1 – 0 –

0 0 – 0 1 1 0 – 1 0 0 0 0 0

0 0 – – 2 0 1 – 2 – – 0 0 0

0 0 – 0 2 0 1 – 2 – – 0 0 0

2 0 – 2 0 0 1 – 2 – – 0 0 2

0 0 – – 0 0 2 – 2 – – 0 0 –

– 2 0 0 0 0 – 0 2 0 0 – – 0

L R L R L L R L L L R R L L

0 0 0 0 2(Hip) 0 – 0 1(TP) 0 – – 0 0

1: regions involved in the epileptogenic zone or EZ (also corresponding to the primary irritative zone); 2: regions involved in the irritative zone or IZ only (secondary irritative zone); 0: regions without electrical abnormalities; –: regions not explored by SEEG; Front Op: fronto-opercular region; Orb: orbital region; PreCing: precingular region; Cing: cingular region; PreF DorsoLat: prefrontal dorsolateral region; PreMot Lat: premotor lateral region; SMA: supplementary motor area; Pre Cent: precentral; Post Cent: postcentral; Ins: Insula; TP: temporal pole; A: Amygdala; T: lateral temporal lobe; Par: parietal lobe; Contr.: contralateral; Hip: hippocampus; L: left; R: right.

the EZ in 8 patients, in the IZ in 9 patients, and outside the EZ and IZ in 13 patients. Therefore, in our study, the sensitivity of diffusion imaging in defining regions that were the site of electrical abnormalities was about 57% for the EZ and 65% for the IZ. However, the specificity was about 7% for both. Differences were observed between patients presenting with visible MR epileptogenic lesion and those with normal appearing MRI. All patients with lesional epilepsy had SID in the EZ and IZ as well as outside leading to a sensitivity of 100% for the EZ and IZ and no specificity. Conversely, 3 out of the 9 patients with negative-MRI had SID in the EZ, 4 in the IZ and 8 outside the EZ and IZ leading to a sensitivity of 33% for the EZ and 44% for the IZ, and a specificity of 11% for both. Figs. 1 and 2 show examples of diffusivity maps, in a negative-MRI patient and in a patient presenting with a post-surgical lesion.

In the 13 patients with diffusion abnormalities present outside the EZ, significant clusters were located in the ipsilateral and in the contralateral frontal lobe in 8 patients (5, 7, 8, 10, 11, 12, 13, 14). All 13 patients exhibited SID outside the frontal lobe. Five patients had only extra-frontal increased diffusivity (1, 2, 4, 6, 9). The extra-frontal diffusion abnormalities were mostly located in the temporal lobe (i.e. all 13 patients with significant increased diffusivity). Basal ganglia were involved by diffusion changes in 7 patients (i.e. putamen in 3 and thalamus in 4) as well as in cerebellum. Six patients had SID in the parietal lobe and 3 in the occipital lobe. At an individual scale (corrected for sampling regions), we also found low Se and Sp. The measures of Se and Sp for each individual subject across sampling points are presented in “Supplementary-Material_Table2” available in the online version.

Fig. 1. Example of significant interictal increased diffusivity in a lesional epilepsy (patient 12). A: Two slices of T1-weighted MRI showing the right frontoorbito-polar post surgical lesion (frontal sinus surgery). B: Significant clusters of increased diffusivity superimposed on diffusivity map. Several clusters are located in the right frontal pole and dorsolateral prefrontal areas, congruent with the location the epileptogenic zone; clusters located in right insulo-opercular areas are congruent with the irritative zone; we note also clusters in remote areas connected to the EZ in the ipsi- and contralateral frontal lobe. R: Right; L: Left.

M. Guye et al. / NeuroImage 35 (2007) 28–37

33

Fig. 2. Example of significant interictal increased diffusivity in a negative-MRI patient (patient 8). A: One slice of T1-weighted normal MRI corresponding to the site of increased diffusivity showed in B. B: Significant clusters of increased diffusivity superimposed on diffusivity map. One cluster located in the mesial part of the left central gyrus (posterior wall of paracentral lobule) is congruent with the location of the epileptogenic confounded with irritative zone in this patient (i.e. central and premotor epilepsy). One can also note the clusters in remote areas connected to the EZ in the ipsilateral and contralateral parietal lobe as well as central and premotor contralateral cortices. R: Right; L: Left.

Concerning the EZ, the overall median was 25% (range 0– 100%) for Se and 43.5% (range 13–100%) for Sp. In the “negativeMRI” group, the median was 0% (range 0–100%) for Se and 53% (range 13–100%) for Sp. In the “positive-MRI” group the median was 100% (range 50–100%) for Se and 40% (range 27–55%) for Sp. Concerning the IZ, the overall median was 31% (range 0– 100%) for Se and 43% (range 13–100%) for Sp. In the “negativeMRI” group, the median was 0% (range 0–43%) for Se and 53% (range 13–100%) for Sp. In the “positive-MRI” group the median was 100% (range 50–100%) for Se and 37% (range 33–53%) for Sp.

of epilepsy (p corrected = 0.026 and Rho corrected = 0.621). Fig. 3 shows this correlation. In addition, SID was significantly more extensive in patients with a visible MR lesion in comparison with negative-MRI patients (p = 0.028). However, no significant difference in extent of diffusion abnormalities was found between groups, in terms of seizure frequency (p = 0.302), secondary generalization (p = 0.841), history of status (p = 0.396), or surgical outcome (p = 0.606). Discussion Methodological issues

Clinical correlations of diffusion abnormalities In our study, a correlation was found between the extent of SID (number of pixels involved by significant changes) and the duration Table 3 Congruence of the location of significant increased diffusivity and the location of EZ and IZ

Negative MRI

MR visible lesion

Patients

EZ

IZ

Outside

1 2 3 4 5 6 7 8 9 10 11 12 13 14

− − − − − − + + + + + + + +

− − − − + − + + + + + + + +

+ + − + + + + + + + + + + +

+ indicate the presence of significant increased diffusivity and − indicate the absence of significant diffusivity changes. EZ: epileptogenic zone; IZ: irritative zone; outside: zones not involved by neither EZ nor IZ.

In the present study we used a short acquisition time DWI sequence (acquired in 144 s), which allowed us to obtain mean diffusivity maps by averaging ADC values recorded in 3 directions (x, y, z). The information provided is the equivalent to the mean diffusivity maps obtained by using DTI. However, DWI does not permit the definition of anisotropy, which is supposed to reflect the asymmetry of water molecule motion. This parameter may also be prone to modification by structural changes. However, previous studies using DTI have shown that diffusivity and anisotropy changes affected different areas and were not usually linked. Moreover, Thivard et al. found anisotropy changes to be more difficult to interpret in the light of depth EEG recordings and discussed almost exclusively the mean diffusivity changes (Thivard et al., 2006). Statistical mapping analysis was chosen to define significant diffusion changes. As in Thivard et al. we did not use the ROI approach to detect quantitative diffusion abnormalities, because of the spatial distribution and the variable location of the EZ and the IZ among our patients (Thivard et al., 2006). Statistical mapping analysis has been extensively used to study brain activation observed by functional imaging (PET or fMRI) (Friston et al., 1995). New application fields of this method now permit the study of variations of morphological parameters such as local percentages and/or local concentrations of grey matter reflecting the extent of local cortical atrophy, referred to as voxel-based-

34

M. Guye et al. / NeuroImage 35 (2007) 28–37

Fig. 3. Correlation between the extent of significant interictal increased diffusivity and epilepsy duration. Increased diffusivity is particularly widespread in patients with an epilepsy duration superior to 18–20 years.

morphometry (VBM) (Friston, 1995). This technique can also be used to compare images from other modalities such as MD (Duncan, 2002a). The advantage of this approach is its ability to compare quantitative parameters derived from imaging modalities throughout the brain, without prior knowledge about the site of pathological change. Nevertheless, several pitfalls can affect the results obtained with this technique. The first problem is the need to perform spatial normalization of data in order to compare parameters from the same regions observed in several subjects. The diffusivity maps were normalized by keeping the value of pixel intensities constant (SPM2) even during spatial transformation (Gitelman et al., 2001). In addition, comparing one subject with a group of controls is a process that can give rise to false positive clusters. We tried to minimize this difficulty by evaluating the extent of diffusivity abnormalities produced by this individual approach, for each control. The robustness of this approach was previously reported on magnetization transfer ratio data (RuggGunn et al., 2003). In the present study, given that only 3 clusters in one control survived the statistical threshold (p < 0.001, FDR corrected p < 0.05) used for the subsequent statistical mapping analysis, we consider that the risk of obtaining a false positive cluster in patients was low. Relationships between diffusion and electrical abnormalities Specificity Our results showed limited specificity of SID in the delineation of the EZ and the IZ. One possible explanation is the discrepancy between the spatial coverage of diffusion imaging and that of SEEG. Indeed SEEG recording implies a spatial sampling based on a priori hypotheses. Theoretically diffusion imaging, having a more extensive sampling area, might detect abnormalities not depicted by SEEG. However, although this could be possible for some interictal abnormalities far beyond the EZ (i.e. secondary irritative zone), this is probably not the case for the EZ. Video-SEEG recording is a method known to allow good prediction of the accuracy of sampling of the EZ. In addition, diffusion abnormalities were found well beyond the EZ as defined by SEEG, in regions unlikely to form part of the EZ in these cases (i.e. occipital lobe, cerebellum and brainstem).

Increased diffusivity areas outside the EZ and IZ were preferentially found in structures known to be functionally connected to the EZ. Firstly, such diffusion changes were more frequently seen in temporal lobe than in parietal and occipital lobes. In this work, 10 of the 14 patients presented with prefrontal epilepsy, 2 of whom had true fronto-temporal epilepsy. As the prefrontal lobe is more directly and functionally connected to the temporal lobe than to the posterior cortices, this could explain the preferential involvement of the temporal lobe in our study. The other areas frequently involved in diffusion changes were also structures strongly connected to the frontal lobe (i.e. contralateral frontal lobe, basal ganglia, cerebellum). A threshold effect is unlikely to explain this relatively low specificity of detection of the EZ, as the most significant clusters were not systematically found in epileptogenic areas (see Table 3). In extra-temporal patients comprising 4 occipital and 2 frontal epilepsies, Thivard et al. found congruent location of SID and EZ in 4 (Thivard et al., 2006). However, in this cohort, only 1 out of the 2 cases of FLE had concordant SID, which is in accordance with our results. Sensitivity In this study, the sensitivity of increased diffusivity in depicting the EZ and the IZ was also relatively low in MRInegative patients but was 100% in lesional epilepsies. Indeed, all MRI-visible lesions exhibited SID, in accordance with the findings of Eriksson et al. in their cohort of MCD (Eriksson et al., 2001). However, SID was never limited to the lesional zone. In addition, we found significantly more widespread areas of increased diffusivity in patients presenting with visible lesions than in those with negative-MRI. These findings suggest a better concordance between SID and structural changes rather than transient functional changes. Potential significance of interictal increased diffusivity in FLE Our data suggest that SID does not simply correspond to causal lesions visible on conventional MRI, but rather reflects structural changes that are not necessarily directly linked to the EZ itself. One possible explanation is that SID may reflect secondary changes related to the propagation of epileptiform discharges within

M. Guye et al. / NeuroImage 35 (2007) 28–37

connected areas. The correlation found between the duration of epilepsy and the spatial extent of increased diffusivity in the present study tends to suggest an evolving process. In addition, as illustrated in Fig. 3, it can be noted that this extent was more widespread in patients whose duration of epilepsy was longer than 18–20 years. This evolution might be secondary to the propagation of epileptiform discharges not only during ictal but also interictal periods, as no correlation between the extent of increased diffusivity and seizure frequency was found. This is corroborated by EEG-correlated functional MRI studies demonstrating hemodynamic responses in regions remote from the epileptogenic areas during both interictal spikes and seizures (Salek-Haddadi et al., 2002, 2006; Kobayashi et al., 2006). These results suggest a widespread rather than a limited effect of spikes and seizures on brain metabolism. However, only a longitudinal study would be able to demonstrate a causal relationship between interictal diffusion changes and epileptiform discharges. Another possible explanation is that the extent of SID might reflect complex modifications of brain structure (potentially arising from different etiological processes) related to the underlying cause of some cases of severe chronic epilepsy. It has previously been shown that, in patients presenting with visible MCD, SID may reflect cortical malformations that are invisible on conventional MRI but detectable by diffusion imaging (Eriksson et al., 2001). In addition, we did not find any effect of the extent of increased diffusivity on eventual surgical outcomes, which also suggests that diffusion abnormalities are not a specific signature of epileptogenic areas. These data suggest that interictal SID is more likely to reflect the consequence of multifactorial structural changes. Pathophysiological hypotheses In animals, acute decrease of water diffusivity during provoked status or sustained seizures reflects a reduction of the extracellular compartment due to cell swelling during cytotoxic edema (Ebisu et al., 1996; Nakasu et al., 1995; Prichard et al., 1995; Zhong et al., 1993, 1997). Diffusivity values returned to normal after a few hours and then frequently increased, transiently (probably due at least in part to vasogenic edema) or permanently (in regions affected by neuronal loss). Thus, in animal models, while the acutely decreased diffusivity seems linked to cellular alterations directly related to sustained epileptiform discharges, secondary permanent increased diffusivity appears as tissue alterations, mainly corresponding to neuronal loss, consecutive to the status. In human studies of status epilepticus, results have also shown decreased diffusivity during the ictal period, but potentially associated with areas of increased diffusivity suggesting more complex osmotic phenomenon between epileptogenic and surrounding areas (Hisano et al., 2000; Wieshmann et al., 1997). Such changes might also reflect the combination of cytotoxic and vasogenic edemas. Recent studies have also demonstrated a persistence of decreased diffusivity in a transient post-ictal period probably due to osmotic changes even in the case of a single seizure (Diehl et al., 2005; Hufnagel et al., 2003; Oh et al., 2004). As in animal studies, human interictal diffusion studies also suggest a potential permanent decreased diffusivity associated with partial epilepsy even in the absence of a clinical history of status epilepticus (Rugg-Gunn et al., 2001). Such an initial event is not required to cause SID. The absence of any association

35

between the extent of SID and a clinical history of status epilepticus in our study tends to corroborate this assumption. Rugg-Gunn et al. showed in a FLE patient that a circumscribed SID, concordant with the EZ in this case, corresponded to diffuse gliosis predominantly of the white matter but no dysgenesis (Rugg-Gunn et al., 2002). Several studies by the same group have led to the hypothesis that increased diffusivity beyond clearly identified MCD is likely to correlate with neuron loss and gliosis, though such changes could also correspond to malformations that are non-visible on conventional MRI (Duncan, 2002a, b). In our study, abnormalities were found in both grey and white matter, suggesting widespread structural changes also concordant with neuronal loss and gliosis. However, neuronal loss and gliosis are not specific of epileptogenic areas, even although they may lead to microscopic structural changes that can affect the neuronal excitability. Our data showed SID not only in the EZ but also in the IZ, as well as in remote areas. Accordingly, previous studies have reported increased T2 relaxation times in regions remote from the EZ, demonstrating structural abnormalities not directly related to the epileptogenic process (Briellmann et al., 2004; Scott et al., 2003). Thus, interictal SID is more likely to reflect multifactorial neuro-glial injuries not specifically related to epileptogenicity. Acknowledgments This work has been supported by a grant from Programme Hospitalier de Recherche Clinique (Ministère de la Santé). M. Guye was supported by a grant from the Assistance PubliqueHopitaux de Marseille and the Centre National de Recherche Scientifique. We thank the neurosurgeons Pr. Peragut and Pr. Dufour. We thank Pr. Figarella-Branger for the Neuropathology.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neuroimage.2006.11.049.

References Arfanakis, K., Hermann, B.P., Rogers, B.P., Carew, J.D., Seidenberg, M., Meyerand, M.E., 2002. Diffusion tensor MRI in temporal lobe epilepsy. Magn. Reson. Imaging 20, 511–519. Ashburner, J., Friston, K.J., 2000. Voxel-based morphometry—The methods. NeuroImage 11, 805–821. Assaf, B.A., Mohamed, F.B., Abou-Khaled, K.J., Williams, J.M., Yazeji, M.S., Haselgrove, J., Faro, S.H., 2003. Diffusion tensor imaging of the hippocampal formation in temporal lobe epilepsy. AJNR Am. J. Neuroradiol. 24, 1857–1862. Bartolomei, F., Wendling, F., Vignal, J.P., Chauvel, P., Liegeois-Chauvel, C., 2002. Neural networks underlying epileptic humming. Epilepsia 43, 1001–1012. Bartolomei, F., Wendling, F., Regis, J., Gavaret, M., Guye, M., Chauvel, P., 2004. Pre-ictal synchronicity in limbic networks of mesial temporal lobe epilepsy. Epilepsy Res. 61, 89–104. Bartolomei, F., Chauvel, P., Wendling, F., 2005. Spatio-temporal dynamics of neuronal networks in partial epilepsy. Rev. Neurol. (Paris) 161 (8–9), 767–780. Basser, P.J., Pierpaoli, C., 1996. Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J. Magn. Reson., B 111, 209–219.

36

M. Guye et al. / NeuroImage 35 (2007) 28–37

Briellmann, R.S., Jackson, G.D., Pell, G.S., Mitchell, L.A., Abbott, D.F., 2004. Structural abnormalities remote from the seizure focus: a study using T2 relaxometry at 3 T. Neurology 63, 2303–2308. Chauvel, P., Trottier, S., Vignal, J.P., Bancaud, J., 1992. Somatomotor seizures of frontal lobe origin. Adv. Neurol. 57, 185–232. Chauvel, P., Kliemann, F., Vignal, J.P., Chodkiewicz, J.P., Talairach, J., Bancaud, J., 1995. The clinical signs and symptoms of frontal lobe seizures. Phenomenology and classification. Adv. Neurol. 66, 115–125 (discussion 125-116). Chauvel, P., Vignal, J.P., Biraben, A., Badier, J.M., Scarabin, J.M., 1996. Stereoelectroencephalography. In: Pavlik, G., Stefan, H. (Eds.), Focus Localization. Multimethodological Assessment of Focalization-Related Epilepsy. Thomas Wiese, Berlin, pp. 135–163. Cossu, M., Cardinale, F., Castana, L., Citterio, A., Francione, S., Tassi, L., Benabid, A.L., Lo Russo, G., 2005. Stereoelectroencephalography in the presurgical evaluation of focal epilepsy: a retrospective analysis of 215 procedures. Neurosurgery 57 (4), 706–718. Diehl, B., Najm, I., Ruggieri, P., Foldvary, N., Mohamed, A., Tkach, J., Morris, H., Barnett, G., Fisher, E., Duda, J., Luders, H.O., 1999. Periictal diffusion-weighted imaging in a case of lesional epilepsy. Epilepsia 40, 1667–1671. Diehl, B., Najm, I., Ruggieri, P., Tkach, J., Mohamed, A., Morris, H., Wyllie, E., Fisher, E., Duda, J., Lieber, M., Bingaman, W., Luders, H.O., 2001. Postictal diffusion-weighted imaging for the localization of focal epileptic areas in temporal lobe epilepsy. Epilepsia 42, 21–28. Diehl, B., Symms, M.R., Boulby, P.A., Salmenpera, T., Wheeler-Kingshott, C.A., Barker, G.J., Duncan, J.S., 2005. Postictal diffusion tensor imaging. Epilepsy Res. 65, 137–146. Duncan, J.S., 2002a. Neuroimaging methods to evaluate the etiology and consequences of epilepsy. Epilepsy Res. 50, 131–140. Duncan, J.S., 2002b. Seizure-induced neuronal injury: human data. Neurology 59, S15–S20. Ebisu, T., Rooney, W.D., Graham, S.H., Mancuso, A., Weiner, M.W., Maudsley, A.A., 1996. MR spectroscopic imaging and diffusionweighted MRI for early detection of kainate-induced status epilepticus in the rat. Magn. Reson. Med. 36, 821–828. Eriksson, S.H., Rugg-Gunn, F.J., Symms, M.R., Barker, G.J., Duncan, J.S., 2001. Diffusion tensor imaging in patients with epilepsy and malformations of cortical development. Brain 124, 617–626. Friston, K.J., 1995. Commentary and opinion: II. Statistical parametric mapping: ontology and current issues. J. Cereb. Blood Flow Metab. 15, 361–370. Friston, K.J., Holmes, A.P., Poline, J.B., Grasby, P.J., Williams, S.C., Frackowiak, R.S., Turner, R., 1995. Analysis of fMRI time-series revisited. NeuroImage 2, 45–53. Gitelman, D.R., Ashburner, J., Friston, K.J., Tyler, L.K., Price, C.J., 2001. Voxel-based morphometry of herpes simplex encephalitis. NeuroImage 13, 623–631. Good, C.D., Johnsrude, I.S., Ashburner, J., Henson, R.N., Friston, K.J., Frackowiak, R.S., 2001. A voxel-based morphometric study of ageing in 465 normal adult human brains. NeuroImage 14, 21–36. Guenot, M., Isnard, J., Ryvlin, P., Fischer, C., Ostrowsky, K., Mauguiere, F., Sindou, M., 2001. Neurophysiological monitoring for epilepsy surgery: the Talairach SEEG method. StereoElectroEncephaloGraphy. Indications, results, complications and therapeutic applications in a series of 100 consecutive cases. Stereotact. Funct. Neurosurg. 77, 29–32. Guye, M., Le Fur, Y., Confort-Gouny, S., Ranjeva, J.P., Bartolomei, F., Regis, J., Raybaud, C.A., Chauvel, P., Cozzone, P.J., 2002. Metabolic and electrophysiological alterations in subtypes of temporal lobe epilepsy: a combined proton magnetic resonance spectroscopic imaging and depth electrodes study. Epilepsia 43, 1197–1209. Guye, M., Ranjeva, J.P., Le Fur, Y., Bartolomei, F., Confort-Gouny, S., Regis, J., Chauvel, P., Cozzone, P.J., 2005. 1H-MRS imaging in intractable frontal lobe epilepsies characterized by depth electrode recording. NeuroImage 26, 1174–1183.

Hisano, T., Ohno, M., Egawa, T., Takano, T., Shimada, M., 2000. Changes in diffusion-weighted MRI after status epilepticus. Pediatr. Neurol. 22, 327–329. Hufnagel, A., Weber, J., Marks, S., Ludwig, T., De Greiff, A., Leonhardt, G., Widmann, G., Stolke, D., Forsting, M., 2003. Brain diffusion after single seizures. Epilepsia 44, 54–63. Jobst, B.C., Williamson, P.D., 2005. Frontal lobe seizures. Psychiatr Clin North Am 28, 635–651, 648–639. Kantarci, K., Shin, C., Britton, J.W., So, E.L., Cascino, G.D., Jack Jr., C.R., 2002. Comparative diagnostic utility of 1 H MRS and DWI in evaluation of temporal lobe epilepsy. Neurology 58, 1745–1753. Kobayashi, E., Bagshaw, A.P., Benar, C.G., Aghakhani, Y., Andermann, F., Dubeau, F., Gotman, J., 2006. Temporal and extratemporal BOLD responses to temporal lobe interictal spikes. Epilepsia 47, 343–354. Lee, J.H., Chung, C.K., Song, I.C., Chang, K.H., Kim, H.J., 2004. Limited utility of interictal apparent diffusion coefficient in the evaluation of hippocampal sclerosis. Acta Neurol. Scand. 110, 53–58. Munari, C., Tassi, L., Di Leo, M., Kahane, P., Hoffmann, D., Francione, S., Quarato, P., 1995. Video-stereo-electroencephalographic investigation of orbitofrontal cortex. Ictal electroclinical patterns. Adv. Neurol. 66, 273–295. Nakasu, Y., Nakasu, S., Morikawa, S., Uemura, S., Inubushi, T., Handa, J., 1995. Diffusion-weighted MR in experimental sustained seizures elicited with kainic acid. AJNR Am. J. Neuroradiol. 16, 1185–1192. Oh, J.B., Lee, S.K., Kim, K.K., Song, I.C., Chang, K.H., 2004. Role of immediate postictal diffusion-weighted MRI in localizing epileptogenic foci of mesial temporal lobe epilepsy and non-lesional neocortical epilepsy. Seizure 13, 509–516. Prichard, J.W., Zhong, J., Petroff, O.A., Gore, J.C., 1995. Diffusionweighted NMR imaging changes caused by electrical activation of the brain. NMR Biomed. 8, 359–364. Rosenow, F., Luders, H., 2001. Presurgical evaluation of epilepsy. Brain 124, 1683–1700. Rugg-Gunn, F.J., Eriksson, S.H., Symms, M.R., Barker, G.J., Duncan, J.S., 2001. Diffusion tensor imaging of cryptogenic and acquired partial epilepsies. Brain 124, 627–636. Rugg-Gunn, F.J., Eriksson, S.H., Symms, M.R., Barker, G.J., Thom, M., Harkness, W., Duncan, J.S., 2002. Diffusion tensor imaging in refractory epilepsy. Lancet 359, 1748–1751. Rugg-Gunn, F.J., Eriksson, S.H., Boulby, P.A., Symms, M.R., Barker, G.J., Duncan, J.S., 2003. Magnetization transfer imaging in focal epilepsy. Neurology 60, 1638–1645. Salek-Haddadi, A., Merschhemke, M., Lemieux, L., Fish, D.R., 2002. Simultaneous EEG-correlated ictal fMRI. NeuroImage 16, 32–40. Salek-Haddadi, A., Diehl, B.;., Hamandi, K., Merschhemke, M., Liston, A., Friston, K., Duncan, J.S., Fish, D.R., Lemieux, L., 2006. Hemodynamic correlates of epileptiform discharges: an EEG-fMRI study of 63 patients with focal epilepsy. Brain Res. 1088 (1), 148–166. Scott, R.C., Cross, J.H., Gadian, D.G., Jackson, G.D., Neville, B.G., Connelly, A., 2003. Abnormalities in hippocampi remote from the seizure focus: a T2 relaxometry study. Brain 126, 1968–1974. Sundgren, P.C., Dong, Q., Gomez-Hassan, D., Mukherji, S.K., Maly, P., Welsh, R., 2004. Diffusion tensor imaging of the brain: review of clinical applications. Neuroradiology 46, 339–350. 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. 1. Introduction and history. Neurochirurgie 20, 1–240. Talairach, J., Bancaud, J., Bonis, A., Szikla, G., Trottier, S., Vignal, J.P., Chauvel, P., Munari, C., Chodkievicz, J.P., 1992. Surgical therapy for frontal epilepsies. Adv. Neurol. 57, 707–732. Thivard, L., Lehericy, S., Krainik, A., Adam, C., Dormont, D., Chiras, J., Baulac, M., Dupont, S., 2005. Diffusion tensor imaging in medial temporal lobe epilepsy with hippocampal sclerosis. NeuroImage 28, 682–690. Thivard, L., Adam, C., Hasboun, D., Clemenceau, S., Dezamis, E., Lehericy, S., Dormont, D., Chiras, J., Baulac, M., Dupont, S., 2006. Interictal

M. Guye et al. / NeuroImage 35 (2007) 28–37 diffusion MRI in partial epilepsies explored with intracerebral electrodes. Brain 129, 375–385. Wang, Y., Majors, A., Najm, I., Xue, M., Comair, Y., Modic, M., Ng, T.C., 1996. Postictal alteration of sodium content and apparent diffusion coefficient in epileptic rat brain induced by kainic acid. Epilepsia 37, 1000–1006. Wieshmann, U.C., Symms, M.R., Shorvon, S.D., 1997. Diffusion changes in status epilepticus. Lancet 350, 493–494.

37

Williamson, P.D., Jobst, B.C., 2000. Frontal lobe epilepsy. Adv. Neurol. 84, 215–242. Zhong, J., Petroff, O.A., Prichard, J.W., Gore, J.C., 1993. Changes in water diffusion and relaxation properties of rat cerebrum during status epilepticus. Magn. Reson. Med. 30, 241–246. Zhong, J., Petroff, O.A., Pleban, L.A., Gore, J.C., Prichard, J.W., 1997. Reversible, reproducible reduction of brain water apparent diffusion coefficient by cortical electroshocks. Magn. Reson. Med. 37, 1–6.