SPM analysis of ictal–interictal SPECT in mesial temporal lobe epilepsy: Relationships between ictal semiology and perfusion changes

Epilepsy Research (2009) 85, 252—260

journal homepage: www.elsevier.com/locate/epilepsyres

SPM analysis of ictal—interictal SPECT in mesial temporal lobe epilepsy: Relationships between ictal semiology and perfusion changes S. Chassagnon a,∗, I.J. Namer b, J.P. Armspach c, A. Nehlig d, P. Kahane e, P. Kehrli f, M.P. Valenti a, E. Hirsch a,g a

Department of Neurology, University Hospital of Strasbourg, France Department of Nuclear Medicine, Strasbourg, France c In vivo Neuroimaging Laboratory, CNRS UMR 7004, Strasbourg, France d Physiopathologie de la schizophrénie, INSERM 666, Strasbourg, France e Department of Neurology, University Hospital of Grenoble, France f Neurosurgery of University Hospital of Strasbourg, France g Centre Thématique de Soins et Recherches-IDEE, Lyon, France b

Received 2 January 2008; received in revised form 8 January 2009; accepted 27 March 2009 Available online 27 June 2009

KEYWORDS Consciousness; Dystonic posturing; Hippocampal sclerosis; Motor automatisms; SPECT; Temporal lobe epilepsy

Summary A combination of temporo-limbic hyperperfusion and extratemporal hypoperfusion was observed during complex partial seizures (CPS) in temporal lobe epilepsy (TLE). To investigate the clinical correlate of perfusion changes in TLE, we analyzed focal seizures of increasing severity using voxel-based analysis of ictal SPECT. We selected 26 pre-operative pairs of ictal—interictal SPECTs from adult mesial TLE patients, seizure-free after surgery. Ictal SPECTs were classified in three groups: motionless seizures (group ML, n = 8), seizures with motor automatisms (MA) without dystonic posturing (DP) (group MA, n = 8), and seizures with DP with or without MA (DP, n = 10). Patients of group ML had simple partial seizures (SPS), while others had CPS. Groups of ictal—interictal SPECT were compared to a control group using statistical parametric mapping (SPM).

Abbreviations: CBF, cerebral blood flow; CPS, complex partial seizures; DLPFC, dorsolateral prefrontal cortex; DP, dystonic posturing; ECD, (99m)Tc]ethyl cysteinate dimmer; EEG, electroencephalography; EZ, epileptogenic zone; MA, motor automatisms; MFG, mid-frontal gyrus; MRI, magnetic resonance imaging; MTLE, mesial temporal lobe epilepsy; MTLS, mesial temporal lobe seizures; MPFC, medial prefrontal cortex; PET, positon emission tomography; PCG, posterior cingulate gyrus; SPECT, single-photon emission computed tomography; SFG, superior frontal gyrus; SISCOM, subtraction ictal single-photon emission computed tomography coregistered to magnetic resonance imaging; SPS, simple partial seizures; TEL, temporal lobe epilepsy. ∗ Corresponding author at: Département de Neurologie, University Hospital of Strasbourg, 1 place de l’hôpital, 67091 Strasbourg Cedex, France. Tel.: +33 388 11 66 62; fax: +33 388 11 63 43. E-mail address: [email protected] (S. Chassagnon). 0920-1211/$ — see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.eplepsyres.2009.03.020

Ictal SPECT in temporal lobe epilepsy

253

In ML group, SPM analysis failed to show significant changes. Hyperperfusion involved the anteromesial temporal region in MA group, and also the insula, posterior putamen and thalamus in DP group. Hypoperfusion was restricted to the posterior cingulate and prefrontal regions in MA group, and involved more widespread associative anterior and posterior regions in DP group. Temporal lobe seizures with DP show the most complex pattern of combined hyper—hypoperfusion, possibly related both to a larger spread and the recruitment of more powerful inhibitory processes. © 2009 Elsevier B.V. All rights reserved.

Introduction SPECT have an established role for peri-ictal imaging in patients with focal drug-resistant epilepsy considered for surgical treatment (Cascino et al., 2004; Devous et al., 1998). In the syndrome of mesial temporal lobe epilepsy (MTLE), the most common type of refractory focal epilepsy, ictal perfusion changes involve not only the ictal onset zone (mesial temporal region), but also cortical and subcortical regions beyond the temporal lobe (Blumenfeld et al., 2004; Tae et al., 2005; Van Paesschen et al., 2003). Hyperperfusion is also associated with ictal hypoperfusion during temporal lobe seizures (Chang et al., 2002; Lee et al., 2000; Tae et al., 2005; Van Paesschen et al., 2003). The underlying mechanism and clinical significance of the latter phenomenon is unclear, although some authors suggested that ictal hypoperfusion could be the consequence of a remote effect arising from subcortical structures (Blumenfeld et al., 2004). More complex patterns of hyperperfusion with the increasing level of complexity and duration of temporal lobe seizures were reported using visual analysis of subtracted SPECT (Kaiboriboon et al., 2005; Shin et al., 2002), but most ictal SPECT studies pooled together various subtypes of complex partial seizures (CPS), without further details about the clinical semiology. Therefore, correlations between ictal blood flow patterns and ictal semiology are still lacking, while MTLE patients may have seizures of various complexities. Although the terms complex partial seizures (CPS) and simple partial seizures (SPS) of the first classification are still widely used (Commission of ILAE, 1981), the task force for epilepsy classification and terminology recently proposed the term ‘‘focal motor seizures with typical (temporal lobe) automatisms’’ to refer to seizures of temporal lobe origin (Engel, 2001). Motor automatisms (MA) and contralateral dystonic posturing (DP) are typical motor symptoms of temporal lobe seizures. Whether and to what extent these symptoms contribute to the complexity of seizures, and therefore reflect different spreading pathways or remote effects, is still unclear. Despite the lack of evidence, MA are believed to result from the functional ‘‘release’’ of extratemporal regions (Loddenkemper and Kotagal, 2005). The pathophysiology of ictal DP is also not fully understood, although neuroimaging clues are pointing to the basal ganglia (Dupont et al., 1998; Mizobuchi et al., 2004). Using ictal SPECT and voxel-based analysis, the aim of our study was to investigate relationships between ictal semiology and ictal blood flow patterns, focusing on the influence

of motor outstanding symptoms of ‘‘typical’’ temporal lobe seizures, i.e. motor automatisms (MA) and dystonic posturing (DP). We hypothesized that the ictal blood flow patterns of CPS with DP versus no DP should be different and that seizures with DP should show more widespread remote effects on extratemporal areas, through the modulatory influence of basal ganglia. To test this hypothesis, we compared ictal SPECTs of temporal lobe seizures, classified according to the presence or absence of MA and DP.

Methods Patients: inclusion criteria We retrospectively reviewed intractable temporal lobe epilepsy (TLE) patients who had ictal and interictal SPECT examinations performed at Strasbourg University Hospital between 1998 and 2002. The inclusion criteria were: (1) adult patients with refractory mesial temporo-limbic seizures (MTLS) according to prolonged interictal and ictal scalp EEG recordings; (2) isolated temporo-limbic sclerosis or hippocampal atrophy on the basis of MR and pathological analysis; (3) early ictal SPECT injection and ongoing seizure activity under video-EEG monitoring during a focal seizure without secondary generalization; (4) interictal SPECT after a seizure-free period of at least 24 h; (5) seizurefree condition (follow-up >4 years) after standardized temporal lobe resection. Patients with bilateral temporal involvement were excluded. Characteristics registered for each patient included clinical history (age at epilepsy onset, duration of epilepsy, seizure frequency and occurrence of occasional secondary generalized seizures), neurological and physical examination, interictal EEG recordings, off-line video-EEG analysis of seizures with detailed description of ictal subjective signs and behavioral changes, neuropsychological assessment, MRI of the brain (including T2-weighted, inversionrecovery T1-weighted, fluid-attenuated inversion recovery).

Ictal EEG recordings and ictal SPECT All patients underwent prolonged video-EEG monitoring with 25 electrodes placed according to the 10—20 system and including 2 additional basal temporal electrodes on each side. For ictal SPECT, EEG technologists trained to radiopharmaceutical administration were assigned to observe patients continuously during periods with a high probability of seizure activity, such as occurs after medication lowering for in-patients, in order to inject the radiotracer (99m)Tc]ethyl cysteinate dimmer (ECD) as soon as possible at the time of the clinical or electrical onset of the seizure (O’Brien, 2000). Patients were asked to warn as soon as possible in case of aura. ECD injection was performed at the earliest time of warning or evidence of motor events or unresponsiveness. Once ECD was injected (that took only a few seconds), the nurse assessed

254 at the bedside responsiveness to simple questions (i.e. ‘‘give me your name’’), ability to perform a simple task in response to a verbal command (i.e. ‘‘close your eyes’’) and asked patients to repeat and remember some objects or numbers to test them for verbal fluency and anterograde and retrograde amnesia. Tonus of the limbs, spontaneous or reactive motor automatisms, vegetative signs (flush, tachycardia, mydriasis and breath changes) were also assessed. Timing of occurrence of clinical events was measured from retrospective off-line analysis of video-EEG recordings, by investigators blinded to the SPECT results. Seizure-onset and seizure-end times were respectively defined as the earliest and latest clinical or electrographic evidence of seizure activity. The injection time of the radiotracer was defined as the time elapsed between the syringe plunger being fully depressed and the seizureonset. Injection time of ECD injection was normalized to seizure duration: normalized injection time (NIT) = [(total seizure duration − ECD injection time)/total seizure duration] × 100. Thus, negative NIT means post-ictal ECD injection, 0 < NIT < 50 means ECD injection after the midpoint of seizure and 50 < NIT < 100 means early ictal ECD injection before the midpoint of the seizure.

Seizure analysis and classification At the time of ictal SPECT, patients were asked to warn as soon as possible in case of aura and were tested for responsiveness, tonus of the limbs, spontaneous or reactive motor automatisms, vegetative signs (flush, tachycardia, mydriasis and breath changes) and asked to remember some objects or numbers to test them for anterograde and retrograde amnesia. Ictal hand or arm DP was defined as sustained unnatural posturing of one upper extremity with tonic and rotational components (Loddenkemper and Kotagal, 2005). MA were defined as involuntary, simple or complex motor movements, resembling normal body movements, often rapid and repetitive, stereotypic or corresponding to the continuation of whatever activity the patient was engaged in, following the arrest of behavior and occurring most often when consciousness is impaired (Kotagal, 1991; Sadler, 2006). For the assessment of consciousness, we used the criteria previously proposed by Lee et al. (2000): complete impairment of consciousness required (inability to follow verbal commands), amnesia of the episode and inability to recall memory items. Consciousness was considered as partially impaired if only one or two items could be evidenced. Ictal SPECTs were then classified in three groups of seizures sharing similar topographic origin but different intensities: Group ML: seizures without motor manifestations; Group MA: seizures with MA and without DP; Group DP: seizures with contralateral arm or hand DP. To allow group analysis, we studied a control group consisting of 10 healthy volunteers with ages between 25 and 40 years.

SPECT data acquisition, pre-processing SPECT studies were acquired with a low-energy, high-resolution double head camera (Elscint Helix, Haifa, Israel) using 740—925 MBq of Technecium-99m-ethyl cysteinate dimer (99mTc-ECD, Neurolite; Du Pont, Wilmington, DE, USA). The camera was operated in the ‘‘stop and shoot’’ mode, with acquisition at three-degree intervals and a total acquisition time of 30 min (120 projections, 64 × 64 matrix). SPECT data were reconstructed using filtered backprojection with a Metz filter. The reconstructed data were post-filtered with an 8 mm full width at half maximum 3D Gaussian filter and displayed in a 128 × 128 matrix. Interictal SPECT studies were performed eyes opened in a quiet environment after a 24-h seizure-free period, under video-EEG monitoring in order to detect subclinical seizures not suitable for interictal reference.

S. Chassagnon et al. Statistical analysis One-way analysis of variance (ANOVA), Neuman—Keuls and Scheffe test were used to search for clinical differences between groups in terms of age at epilepsy onset, epilepsy duration, frequency of seizures, seizure duration during the ictal SPECT, time for tracer injection and age at surgery. Statistical comparison of each group of pairs of interictal/ictal SPECT with the control group was performed by multi-group analysis using the statistical parametric mapping (SPM) software (Wellcome Department of Cognitive Neurology, Institute of Neurology, University College London, UK), with the SPM2 version implemented in MATLAB (Mathworks Inc., Sherborn, MA, USA). The image of the right and left patients were combined for analysis by flipping right images to show all changes on the left side, in order to align the epileptogenic zone on the same side. All scans were normalized to the same stereotaxic space. A voxel-by-voxel comparison according to the general linear model and t statistics was used to calculate the difference in cerebral perfusion between the two conditions (ictal versus interictal) in each group of pairs of ictal/interictal SPECTs, and the differences between two conditions (rest versus rest) of control group. Comparisons were made using non-paired two-sample t-test. In order to control for familywise errors due to multiple comparisons, the height threshold was set to false discovery rate (FDR) with a corrected P value of less than 0.01 and the extent threshold was set to k > 115 voxels. The cluster size was set to correspond to the level of spatial resolution of our SPECT system (about 12 mm). SPM images were then displayed in three orthogonal planes by using a ‘‘glass brain’’ and the regions of significance were overlaid onto a template MRI to define the anatomic regions of perfusion changes by using their x, y, z coordinates in the Talairach space. We matched for each comparison the number of pairs of SPECTs from healthy individuals and patients.

Results Clinical data Based on the inclusion criteria, 24 patients with MTLE (12 left, 12 right) were enrolled, among 150 patients with focal drug resistant epilepsy who underwent ictal and interictal SPECT during their pre-surgical workup. The mean age at onset of epilepsy was 14 years (range, 1—34), the mean epilepsy duration was 21 years (range, 3—48) and mean frequency of seizures before surgery was 6 per month (range, 2—20). Three patients underwent depth electrode recordings in Grenoble (France), because of incongruent clinical data (suspicion of bilateral involvement, n = 1; suspicion of early temporo-frontal involvement, n = 2). A temporo-mesial onset was also confirmed in these three patients, who were successfully treated by a standardized temporal resection. All but two patients are seizure-free with a mean follow-up after surgery of 78 months (range, 61—111). The two remaining patients have had rare seizures after surgery (Engel class IIa, n = 1; class IIb, n = 1). Pathological analysis confirmed the presence of a typical mesial sclerosis in 21 patients; the others had a mild neuronal loss in Ammon’s horn without gliotic changes (n = 1), minor microdysgenesic changes in the anterior temporal white matter (n = 1) and no reliable results due to damaged brain sample (n = 1). Among a total of 160 seizures recorded during prolonged scalp-EEG recordings (range, 3—14 per patient), we focused our study on the focal seizures recorded during 26 ictal

0 0 0 0 3 2 0 0 10 0 8 7

Consciousness was partially impaired in the three remaining patients of the group MA.

0 5a 10 2 2 3 8 5 8 78 (41—89) 72 (48—89) 79 (59—90) 11 (3—22) 30 (7—61) 16 (9—31)

Secondary generalization Head version Dystonic posturing Motor automatisms Oro-alimentary automatisms Impairment of consciousness Vegetative symptoms Aura Normalized injection time SPECT injection time (s)

0 5 8 a

The results of the SPM analysis are presented in Table 2 and Fig. 1. Group ML did not differ from the control group. Nevertheless, a cluster of hypoperfused voxels could be observed in the intermediate part of the contralateral frontal lobe (x = 12, y = 28, z = 20) when the extent threshold was lowered to 50 voxels (k = 50, Z = 3.6).

ML (n = 8) 50 (13—114) MA (n = 8) 104 (30—180) DP (n = 10) 82 (47—104)

Statistical analysis of perfusion changes

Seizure duration (s)

SPECTs from 24 patients, since two patients of group ML also had ictal SPECT during CPS, that were included in the MA (n = 1) and DP (n = 1) groups. A total of 26 ictal SPECT were then available for group comparisons: group ML = 8 (3 women, 5 men), group AU = 8 (4 women, 4 men) and group DP = 10 (8 women, 2 men). In all the eight patients of group ML, ictal symptoms were similar to those usually reported by these patients at the onset of complex partial seizures. ML seizures corresponded most of the time to auras, with behavioral arrest and moderate vegetative signs (slight tachycardia and/or rubbing). The epileptic nature of these events was also supported by objective ictal EEG changes. Off-line analysis of ictal videotapes and EEGs confirmed that the seizure recorded at the time of the ictal SPECT was representative of the patient’s usual seizures and showed at least one of the four features: (1) clear-cut lateralized attenuation of the alpha rhythm, (2) disappearance of interictal temporal lobe epileptiform abnormalities, (3) a focal flattening over temporal lobe electrodes and (4) a rhythmic transient theta activity over temporal lobe electrodes. Lateralization of ictal EEG changes was always congruent with the side of hippocampal sclerosis. 5/8 patients of group ML, 3/8 patients of group MA and 6/10 patients of group DP had a right temporal lobe epilepsy. Consciousness was partially or totally impaired in 9/14 patients with right temporal lobe epilepsy and 9/12 patients with left temporal lobe epilepsy. In addition, 3/8 patients of group ML had left-sided TLE with preserved responsiveness. Among the three patients in group MA with partial impairment of consciousness, 2/3 had left temporal lobe epilepsy and 1/3 had right temporal lobe epilepsy. Seizures with loss of consciousness and automatisms were also recorded in patients of group ML but were not imaged in 6/8 of them by ictal SPECT. Timing of injection and outstanding ictal symptoms assessed at the time of the ictal SPECT are summarized in Table 1. Motor automatisms were bilateral in 6/8 patients of group MA and 2/10 patients of group DP, or affected only the upper arm ipsilaterally to the hemisphere of seizure onset in 2/8 patients of group MA, 5/10 patients of group DP. Careful ictal assessment of consciousness and off-line video analysis showed that all patients of group ML had SPS (no impairment of consciousness), while patients of groups MA and DP had CPS with complete (n = 15) or partial loss of consciousness (n = 3). There were no differences among groups concerning the age at epilepsy onset, epilepsy duration, frequency of seizures. Normalized injection times showed true ictal injection time in all patients, before the midpoint of seizures in all cases but two. Nevertheless, the mean seizure duration (seizure recorded during the ictal SPECT) and injection time of tracer were significantly shorter (p < 0.05) in group ML in comparison with group MA.

255 Table 1 Ictal semiology in the three groups of patients at the time of the ictal SPECT. Median (and range) of seizure duration and injection times are expressed in seconds. Normalized injection time = [(seizure duration − time of injection)/seizure duration] × 100. The numbers indicate the number of patients in each group having the concerning symptoms.

Ictal SPECT in temporal lobe epilepsy

256 Table 2 Ictal perfusion changes during mesial temporal lobe seizures: t-test, SPM analysis. Hyperperfusion and hypoperfusion were evaluated separately by t-test and appropriate contrasts. We used an extent threshold k = 115 voxels, voxel-level significance threshold p = 0.001, followed by a cluster-level significance level p = 0.05 corrected for multiple comparisons for the entire brain. Groups: ML: motionless seizures; MA: seizures with MA and without dystonic posturing, DP: seizures with contralateral arm dystonic posturing. Talairach coordinates [x, y, z] are expressed in mm; FDR < 0.05 (false rate discovery); k: number of voxel per cluster; Z = significance. Abbreviations: ACG: anterior cingulate Gyrus, BA: Brodmann area, Cx: cortex, IPL: inferior parietal lobule, ITG: inferior temporal gyrus, MFG: mid-frontal gyrus, MTG: mid-temporal gyrus, PCG: posterior cingulate gyrus, PFC: prefrontal cortex, PHG: para-hippocampal gyrus, SFS: superior frontal sulcus, SMA: supplementary motor area and TP: temporal pole. Hyperperfusion x

y

Hyperperfusion z

ML

—

—

—

MA

−28 −52 −35

−6 16 8

−30 −20 −40

DP

−10 −30 −46

−20 −2 18

a b

12 −16 −26

Cluster size (k)

Peak Z

Anatomic localization

—

—

—

2382

3339

5.62 4.49 4.07 5.09 4.49 3.91

x

a

Entorhinal Cx, BA 28/36 TP (supero-lateral), BA 38a TP (inferior), BA 38a Mediodorsal thalamus Amygdala TP (apex), BA 38

a

y

z

—

—

—

8 −2 38

—52 −52 46

22 24 20

−30 −6 −24 −16 −6 −8 −34 8 10 −42 −48 −52 30

42 24 18 48 50 30 48 40 42 −58 −56 −50 −68

10 54 44 6 12 34 −10 22 −28 26 0 −10 −24

Cluster size (k)

Peak Z

Anatomic localization

—

—

—

3.86 3.61 3.8

PCG, BA 31b PCG, BA 31a PFC, BA 9/46b

4.49 4.25 4.12 4.08 4.02 3.99 3.89 4.44 3.85 3.99 3.49 3.12 4.32

Anterior MFG, BA 46/10a MPFC, BA 6/8 (pre-SMA)a PFC overlapping SFS, BA 8a Frontal pole, BA 32/10a MPFC, BA 10a ACG, BA 32a Orbito-frontal Cx, BA 47/10a ACG, BA 32b Gyrus rectus, BA 11b IPL, BA 39a Posterior MTG, BA 21/37a Posterior ITG, BA 37a Lateral cerebellumb

486 214

8471

828 1130

Ipsilateral to seizure focus. Contralateral.

S. Chassagnon et al.

Ictal SPECT in temporal lobe epilepsy

257

Figure 1 Brain areas with significant perfusion changes during complex partial seizures in patients with MTLE, shown on a single brain MRI scan. Results of t-test are shown on sections A—C. A: coronal sections progressing from anterior to posterior; B: sagittal sections from the lateral aspect of the epileptic hemisphere to the midline; C: horizontal sections from inferior to superior. Ictal hyperperfusion is in yellow for group MA (n = 8), red for group DP (n = 10); overlap between groups MA and DP appears in orange; ictal hypoperfusion is in green for group MA and blue for group DP. The image of the right and left patients were combined by flipping right images. Left is on the left. Hyperperfusion was found in anterior mesial temporal regions in both groups, extending in the postero-inferior part of insular cortex, the posterior putamen and the mediodorsal thalamus in group DP. In group MA, ictal hypoperfusion was observed in lateral prefrontal cortex contralaterally and posterior cingulate cortex. In group DP, hypoperfusion involved bilaterally posterior cingulate cortex and orbitofrontal cortex with ipsilateral predominance, ipsilaterally medial prefrontal cortex and temporo—parieto—occipital junction, contralaterally the cerebellum. In each group, ictal—interictal pairs of SPECTs were compared with the same number of pairs of SPECTs from healthy normal individuals using statistical parametric mapping (SPM 2002). Extent threshold, k = 115 voxels (equivalent to a sphere of 900 mm3 with respect to the 12 mm spatial resolution of the camera, taking into account the SPM voxel size of 2 mm × 2 mm × 2 mm), height threshold, p < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Hyperperfusion involved temporal antero-medial structures in groups MA and DP, also extending to the insula, posterior part of the putamen and thalamus in the group DP. In groups MA and DP, hypoperfusion involved anterior frontal and mesial frontal regions, the posterior cingulate cortex (PCC), mid-frontal gyri, frontal pole and supramarginal gyrus. Group DP also revealed hypoperfusion in parieto-occipital and temporo-occipital junctions, the posterior part of the para-hippocampal gyrus, the inferior parietal lobule and posterior temporal region. Seizures with DP showed the more widespread pattern of combined hyper/hypoperfusion.

Discussion Using ictal SPECT and voxel-based analysis in MTLE, we observed composite patterns of hyper- and hypoperfusion during mesiotemporal CPS. CPS with DP showed more widespread pattern of hyper- and hypoperfusion than CPS without DP. We suggest that DP reflects both a more widespread propagation of temporo-limbic discharges (insula, thalamus and basal ganglia) and also a more intense remote extratemporal deactivation through the modulatory influence of subcortical structures, i.e. basal ganglia. Our study is the first one using voxel-based analysis to show distinct patterns of perfusion among so-called ‘‘complex partial seizures’’ of temporal lobe origin.

258

Methodological concerns The limitation of blood flow studies concerns the shortage of knowledge about neuronal activity that generates increases or decreases of blood flow. Epileptic discharges, responsible for increases of neuronal firing and synaptic activity, are the main contributors to blood flow increases, without specificity for particular types of neurotransmission (Logothetis et al., 2001), and the hemodynamic signals accompanying a functional activation is mainly dependent on the afferent function (Lauritzen, 2001). On the other hand, ictal hypoperfusion results from different mechanisms, including at least: (1) an active trans-synaptic inhibition arising from the epileptic focus, i.e. ‘‘ictal surround inhibition’’ related to the inhibitory collateral effects of interneurons (Schwartz and Bonhoeffer, 2001); (2) the deactivation of neural networks with respect to their baseline activity at rest, as a consequence of the remote effect of epileptic discharges (Shulman et al., 1997; Gotman et al., 2005). A possible limitation of the present study was the small number of patients and the fact that we pooled together patients with right and left focus. Asymmetries in the ictal pattern of perfusion between right-sided and left-sided TLE were previously reported in the temporo-parietal junction and brainstem, but were not observed in the temporal lobe, thalamus and basal ganglia (Tae et al., 2005; Hogan et al., 2006).

S. Chassagnon et al. nisms, i.e. different mode of spreading of temporo-limbic discharges. We observed significant hyperperfusion in the posterior part of the putamen and the lateral pallidum ipsilaterally in group DP. Interictal hypometabolism and ictal hyperperfusion or of the basal ganglia was observed more frequently during temporal lobe seizures with DP, in comparison with seizures without DP (Dupont et al., 1998; Mizobuchi et al., 2004; Joo et al., 2004). However, ictal SPECT studies mixed SPS and CPS in the group of seizures without DP that was not the case in our study, where we tried to get groups as homogeneous as possible with respect to ictal phenomenology. The pathophysiology of ictal DP is not fully understood. Animal studies provided evidence for the involvement of the basal ganglia in the remote inhibitory control of seizures (Deransart and Depaulis, 2002). In humans, Ictal DP was correlated with ictal scalp EEG slowing (Rusu et al., 2005; Kim et al., 2007), and EEG slowing within the putamen using depth recordings (Kuba et al., 2003). The presence of MA during MTLS is also associated with ictal scalp EEG slowing, over fronto-temporal regions (Rusu et al., 2005). The reason why we did not find extratemporal hyperperfusion in the group MA is possibly related to the small number of patients. Moreover, the group analysis only shows generic features of ictal blood flow changes, smoothing what is related to the inter-individual variability.

Ictal patterns of hyperperfusion

Ictal hypoperfusion and combined patterns of hyper- and hypoperfusion

Group analysis failed to show significant perfusion changes in ML group, either within the temporal lobe or outside. Blumenfeld et al. (2004) also failed to detect significant changes in six patients with SPS but injected the tracer postictally. This was not the case in the present study in which the mean injection time was 11 s after seizure onset. However, regarding the relatively short duration of seizures in this group (50 s), and taking into account the biodistribution of ECD in humans, we assume that our results reflect a mixture of ictal and early post-ictal events. In group DP only, but not in group MA, hyperperfusion involved extratemporal regions, including the insula and subcortical structures. Hyperperfusion of the insula was already reported in TLE patients (Shin et al., 2002; Tae et al., 2005) and the role of the insula in the spreading of temporal lobe seizures was confirmed using depth recordings (Isnard et al., 2000). Ipsilateral or bilateral thalamic hyperperfusion was also reported in some ictal SPECT studies (Blumenfeld et al., 2004; Joo et al., 2004; Tae et al., 2005) but not in others (Chang et al., 2002; Van Paesschen et al., 2003). Evidence coming from depth recordings supports the role of the thalamus in the spreading of temporal lobe seizures (Rosenberg et al., 2006). With respect to ictal phenomenology, hyperperfusion of the thalamus during CPS was correlated with the impairment of consciousness (Lee et al., 2002; Blumenfeld and Taylor, 2003), but these authors pooled different type of CPS. We did not observe thalamic hyperperfusion in CPS without DP, despite complete (5/8) or partial (3/8) impairment of consciousness in this group. Therefore, we suggest that thalamic CBF changes are not only related to the level of consciousness, but could also reflect different mecha-

Hypoperfusion involved the PCC in both groups of CPS, as well as frontal (group MA and DP) and posterior cortical regions (group DP). Combined patterns of hyper- and hypoperfusion were previously reported during temporal lobe seizures (Menzel et al., 1998) and confirmed using group analysis (Chang et al., 2002; Van Paesschen et al., 2003; Amorim et al., 2005; Tae et al., 2005; Nelissen et al., 2006), but ictal semiology was not detailed. In these studies, hypoperfusion involved a set of associative cortical regions, including prefrontal regions, PCG and ACG, inferior and posterior temporal gyri, paracentral lobule (bilaterally) and supramarginal gyrus (ipsilaterally). A significant correlation between ictal hyperperfusion of the temporal lobe and ictal hypoperfusion of the frontal lobe was statistically demonstrated in patients with CPS (Van Paesschen et al., 2003). Since hypoperfusion involved fronto-parietal and posterior cingulate areas, one could argue that this pattern shared some similarity with the ‘‘default state network of the brain’’, a set of cortical areas tonically active in resting physiological conditions and deactivated (BOLD decrease or relative hypometabolism) during various visual and verbal tasks (Shulman et al., 1997). Moreover, Gotman et al. (2005) also observed BOLD decreases in a set of cortical areas that matched the ‘‘default state’’ during generalized interictal discharges. More recently, Laufs et al. (2006) reported a bilateral deactivation of the PCC, precuneus, anterior frontal and inferior parietal regions during interictal epileptiform discharges in nine patients with left temporal focus, using EEG-fMRI. We assume that the suspension of the default state of the brain can not explain all the CBF

Ictal SPECT in temporal lobe epilepsy decreases observed in our study for two main reasons: (1) hypoperfusion was more widespread during CPS with DP than CPS without DP, extending beyond the so-called ‘‘default state network’’ and (2) asymmetry of hypoperfusion, mainly in the group DP. Although the frontal lobe regions involved in the default mode are mainly ventro—anteromedial and anterior cingulate regions (Shulman et al., 1997), we also observed hypoperfusion in the dorsolateral and dorsomedial prefrontal cortices during CPS. DP could reflect both spreading to subcortical structures, and their recruitment as a circuit control against seizures, thus leading to more pronounced and widespread deactivation of extratemporal cortical areas.

Is dystonic posturing a sign of seizure complexity? Complex partial seizures arise most commonly from the temporal lobe and are usually defined as focal seizures with impairment of consciousness. The current conception is that the complexity of focal seizures reflects the involvement of both mesial and lateral temporal lobe structures, or temporal lateral onset or widespread bilateral cortical areas (Maillard et al., 2004). On the other hand, the ‘‘network inhibition hypothesis’’ attaches importance to seizure spread to midline subcortical structures, leading in turn to bilateral cortical inhibition (Norden and Blumenfeld, 2002). The latter authors also observed a correlation between the occurrence of the impairment of consciousness and the combined pattern of thalamic hyperperfusion and fronto-parietal hypoperfusion at the group level (Blumenfeld et al., 2004). However, 8/13 (complex partial seizures of any origin) and 1/9 (TLE) patients had a complete or partial preserved responsiveness. More recently, electrophysiological studies of intracerebral synchrony suggest that the thalamic recruitment could mainly take place in the late ictal or even early post-ictal period, at the time when inhibitory processes concur to stop the seizure (Guye et al., 2006). The contribution of ictal motor symptoms to the complexity of combined patterns of hyper- and hypoperfusion has not yet been investigated. Although some patients with CPS and DP showed more widespread ictal hyperperfusion (Shin et al., 2002; Joo et al., 2004) or interictal hypometabolism (Rusu et al., 2005) than CPS without DP, the influence of DP on CBF patterns was never studied using group analysis and ictal hypoperfusion was never taken into account. Rusu et al. (2005) stated that patients with ictal DP had more severe epilepsy than those without DP, since they had longer seizure duration, more complex ictal semiology, higher frequency of secondary generalization, severe clouding of consciousness, and prolonged post-ictal confusion. Although we did not assess the duration of post-ictal confusion and keeping in mind that three patients of group MA had partial impairment on consciousness, there was no major differences between CPS with and without DP indicating that CPS with DP were more severe. Thus, comparing the CBF patterns of the two group of CPS, we suggest that DP seems to be per se a marker of seizure severity. Recently, a study of correlation between EEG and CBF patterns in CPS showed

259 that the more widespread ictal scalp EEG pattern was found during CPS with DP (Kim et al., 2007). The co-occurrence of thalamic and basal ganglia hyperperfusion at the time of DP is in line with the role of some compartments of these structures in the remote control of temporal lobe seizures in animal studies (Cassidy and Gale, 1998; Deransart and Depaulis, 2002). Finally, the more widespread hypoperfusion related to CPS with DP versus no DP, could reflect protective mechanisms, recruited to inhibit or interrupt seizures.

Conclusion Dystonic posturing could be per se a marker of seizure severity in term of perfusion changes, since mesiotemporal lobe seizures with contralateral dystonic posturing shows more widespread hyper- and hypoperfusion than complex temporal seizures without dystonic posturing. The spatial extent and distribution of ictal hypoperfusion depends both on spreading and remote effect of mesiotemporal lobe originating discharges.

Conflict of interest The authors report no conflicts of interest.

Acknowledgements We thank Pr Meyer (Department of Statistics, Strasbourg) and Pr I.J. Namer (Department of Nuclear Medicine, Strasbourg) for their assistance in statistical analysis of the data, the staph of neurology for the patients’ assessment and data collection, Pr Gotman, Drs. Dubeau and Grova (Montreal) for their critical contribution to the manuscript.

References Amorim, B.J., Etchebehere, E.L., Sa de Camargo, Camargo, E.E., Rio, P.A., Bonilha, L., Rorden, C., Li, M.L., Cendes, F., 2005. Statistical voxel-wise analysis of ictal SPECT reveals pattern of abnormal perfusion in patients with temporal lobe epilepsy? Arq. Neuropsiquiatr. 6 (4), 977—983. Blumenfeld, H., Taylor, J., 2003. Why do seizures cause loss of consciousness? Neuroscientist 9 (5), 301—310. Blumenfeld, H., McNally, K.A., Vanderhill, S.D., Paige, A.L., Chung, R., Davis, K., Norden, A.D., Stokking, R., Studholme, C., Novotny Jr., E.J., Zubal, I.G., Spencer, S.S., 2004. Positive and negative network correlations in temporal lobe epilepsy. Cereb. Cort. 14, 892—902. Cascino, G.D., So, E., Buchhalter, J.R., Mullan, B.P., 2004. The current place of single photon emission computed tomography in epilepsy evaluations. Neuroimaging Clin. N. Am. 14 (3), 553—561. Cassidy, R.M., Gale, K., 1998. Mediodorsal thalamus plays a critical role in the development of limbic motor seizures. J Neurosci. 18 (21), 9002—9009. Chang, D.J., Zubal, I.G., Gottschalk, C., Necochea, A., Stokking, R., Studholme, C., Corsi, M., Slawski, J., Spencer, S.S., Blumenfeld, H., 2002. Comparison of statistical parametric mapping and SPECT difference imaging in patients with temporal lobe epilepsy. Epilepsia 43 (1), 68—74. Commission on Classification Terminology of the International League Against Epilepsy, 1981. Proposal for revised clinical

260 and electroencephalographic classification of epileptic seizures. Epilepsia 22 (4), 489—501. Deransart, C., Depaulis, A., 2002. The control of seizures by the basal ganglia? A review of experimental data. Epileptic Disord. 4 (Suppl. 3), S61—72. Devous Sr., M.D., Thisted, R.A., Morgan, G.F., Leroy, R.F., Rowe, C.C., 1998. SPECT brain imaging in epilepsy: a meta-analysis. J. Nucl. Med. 39 (2), 285—293. Dupont, S., Semah, F., Baulac, M., Samson, Y., 1998. The underlying pathophysiology of ictal dystonia in temporal lobe epilepsy: an FDG-PET study. Neurology 51 (5), 1289—1292. Engel Jr., J., 2001. A proposed diagnostic scheme for people with epileptic seizures and with epilepsy: report of the ILAE task force on classification and terminology. Epilepsia 42, 796— 803. Gotman, J., Grova, C., Bagshaw, A., Kobayashi, E., Aghakhani, Y., Dubeau, F., 2005. Generalized epileptic discharges show thalamocortical activation and suspension of the default state of the brain. Proc. Natl. Acad. Sci. 102 (42), 15236—15240. Guye, M., Regis, J., Tamura, M., Wendling, F., McGonigal, A., Chauvel, P., Bartolomei, F., 2006. The role of corticothalamic coupling in human temporal lobe epilepsy. Brain 129, 1917— 1928. Hogan, R.E., Kaiboriboon, K., Bertrand, M.E., Rao, V., Acharya, J., 2006. Composite SISCOM perfusion patterns in right and left temporal seizures. Arch. Neurol. 63, 1419—1426. Isnard, J., Guenot, M., Ostrowsky, K., Sindou, M., Mauguiere, F., 2000. The role of the insular cortex in temporal lobe epilepsy. Ann. Neurol. 48 (4), 614—623. Joo, E.Y., Hong, S.B., Lee, E.K., Tae, W.S., Kim, J.H., Seo, D.W., 2004. Regional cerebral hyperperfusion with ictal dystonic posturing: ictal—interictal SPECT subtraction. Epilepsia 45 (6), 686—689. Kaiboriboon, K., Bertrand, M.E., Osman, M.M., Hogan, R.E., 2005. Quantitative analysis of cerebral blood flow patterns in mesial temporal lobe epilepsy using composite SISCOM. J. Nucl. Med. 46, 38—43. Kim, J.H., Im, K.C., Kim, J.S., Lee, S.A., Lee, J.K., Khang, S.K., Kang, J.K., 2007. Ictal hyperperfusion patterns in relation to ictal scalp EEG patterns in patients with unilateral hippocampal sclerosis: a SPECT study. Epilepsia 48 (2), 270—277. Kotagal, P., 1991. Seizure symptomatology of temporal lobe epilepsy. In: Luders, H.O. (Ed.), Epilepsy surgery. Raven Press Ltd., New York, pp. 143—156. Kuba, R., Rektor, I., Brazdil, M., 2003. Ictal limb dystonia in temporal lobe epilepsy. An invasive video-EEG finding. Eur. J. Neurol. 10 (6), 641—649. Laufs, H., Hamandi, K., Salek-Haddadi, A., Kleinschmidt, A.K., Duncan, J.S., Lemieux, L., 2006. Temporal lobe interictal epileptic discharges affect cerebral activity in ‘‘default mode’’ brain regions. Hum. Brain Mapp. 27 (11). Lauritzen, M., 2001. Relationship of spikes, synatic activity, and local changes of cerebral blood flow. J. Cereb. Blood Flow Metab. 21, 1367—1383. Lee, H.W., Hong, S.B., Tae, W.S., 2000. Opposite ictal perfusion patterns of subtracted SPECT. Hyperperfusion and hypoperfusion. Brain 123 (10), 2150—2159.

S. Chassagnon et al. Lee, K.H., Meador, K.J., Park, Y.D., King, D.W., Murro, A.M., Pillai, J.J., Kaminski, R.J., 2002. Pathophysiology of altered consciousness during seizures: Subtraction SPECT study. Neurology 59 (6), 841—846. Loddenkemper, T., Kotagal, P., 2005. Lateralizing signs during seizures in focal epilepsy. Epilepsy Behav. 7, 1—17. Logothetis, N.K., Pauls, J., Augath, M., Trinath, T., Oeltermann, A., 2001. Neurophysiological investigation of the basis of the fMRI signal. Nature 412 (6843), 150—157. Maillard, L., Vignal, J.P., Gavaret, M., Guye, M., Biraben, A., McGonigal, A., Chauvel, P., Bartolomei, F., 2004. Semiologic and electrophysiologic correlations in temporal lobe seizure subtypes. Epilepsia 45 (12), 1590—1599. Menzel, C., Grunwald, F., Klemm, E., Ruhlmann, J., Elger, C.E., Biersack, H.J., 1998. Inhibitory effects of mesial temporal partial seizures onto frontal neocortical structures. Acta Neurol. Belg. 98 (4), 327—331. Mizobuchi, M., Matsuda, K., Inoue, Y., Sako, K., Sumi, Y., Chitoku, S., Tsumaki, K., Takahashi, M., 2004. Dystonic posturing associated with putaminal hyperperfusion depicted on subtraction SPECT. Epilepsia 45 (6), 686—689. Nelissen, N., Van Paesschen, W., Baete, K., Van Laere, K., Palmini, A., Van Billoen, H., Dupont, P., 2006. Correlations of interictal FDG-PET metabolism and ictal SPECT perfusion changes in human temporal lobe epilepsy with hippocampal sclerosis. Neuroimage 32 (2), 684—695. Norden, A.D., Blumenfeld, H., 2002. The role of subcortical structures in human epilepsy. Epilepsy Behav. 3 (3), 219—231. O’Brien, T.J., 2000. SPECT: methodology. Adv. Neurol. 83, 11—32. Rosenberg, D.S., Mauguiere, F., Demarquay, G., 2006. Involvement of medial pulvinar thalamic nucleus in human temporal lobe seizures. Epilepsia 47 (1), 98—107. Rusu, V., Chassoux, F., Landre, E., Bouilleret, V., Nataf, F., Devaux, B.C., Turak, B., Semah, F., 2005. Dystonic posturing in seizures of mesial temporal origin: electroclinical and metabolic patterns. Neurology 65 (10), 1612—1619. Sadler, R.M., 2006. The syndrome of mesial temporal lobe epilepsy with hippocampal sclerosis: clinical features and differential diagnosis. Adv. Neurol. 97, 27—37. Schwartz, T.H., Bonhoeffer, T., 2001. In vivo optical mapping of epileptic foci and surround inhibition in ferret cerebral cortex. Nat. Med. 7 (9), 1063—1067. Shin, W.C., Hong, S.B., Tae, W.S., Kim, S.E., 2002. Ictal hyperperfusion patterns according to the progression of temporal lobe seizures. Neurology 58, 373—380. Shulman, G.L., Fiez, J.A., Corbetta, M., Buckner, R.L., Miezin, F.M., Raichle, M.E., 1997. Common blood flow changes across visual tasks. II. Decreases in cerebral cortex. J. Cogn. Neurosci. 19, 648—663. Tae, W.S., Joo, E.Y., Kim, J.H., Han, S.J., Suh, Y.L., Kim, B.T., Hong, S.C., Hong, S.B., 2005. Cerebral perfusion changes in mesial temporal lobe epilepsy: SPM analysis of ictal and interictal SPECT. Neuroimage 24 (1), 101—110. Van Paesschen, W., Dupont, P., Van Driel, G., Van Billoen, H., Maes, A., 2003. SPECT perfusion changes during complex partial seizures in patients with hippocampal sclerosis. Brain 126, 1103—1111.