EEG–fMRI in focal epilepsy: Local activation and regional networks

Clinical Neurophysiology 125 (2014) 21–31

Contents lists available at SciVerse ScienceDirect

Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph

EEG–fMRI in focal epilepsy: Local activation and regional networks D. Flanagan a, R.A.B. Badawy a,b,c, G.D. Jackson a,b,c,⇑ a

Brain Research Institute, Florey Institute of Neuroscience and Mental Health, Victoria, Australia Department of Neurology Austin Health, Heidelberg, Australia c Epilepsy Research Centre, Department of Medicine, University of Melbourne, Melbourne, Australia b

a r t i c l e

i n f o

Article history: Accepted 27 June 2013 Available online 17 July 2013 Keywords: EEG–fMRI IED Focal epilepsy Piriform cortex

h i g h l i g h t s  EEG/fMRI is an important new tool for studying patients with focal epilepsy and the electrographic

field of the interictal epileptiform discharges (IEDs) is reflected in the fMRI results with focal and diffuse IEDs providing novel and important localization information.  Amongst patients with a heterogeneous array of focal epilepsies, the piriform cortex is a common node in the underlying networks associated with focal IEDs.  In cases of diffuse IEDs we noted involvement of subcortical structures, in particular the thalamus and cerebellum.

a b s t r a c t Objective: To identify features of BOLD signal change associated with interictal epileptiform discharges (IEDs) in a heterogeneous group of focal epilepsy patients. Methods: EEG/fMRI studies in 27 focal epilepsy patients were reviewed with attention given to the extent and location of the IED and the resulting pattern of BOLD signal change. Second order group analysis was used to identify common features. Results: fMRI results provided novel clinical information for individual patients. We identified a significant common node within the ipsilateral piriform cortex as well as patterns involving distant cortical or subcortical areas. Conclusion: Despite the heterogeneity of IEDs in focal epilepsy, there are important common features underpining IEDs with a highly significant fMRI node in the ipsilateral piriform cortex. Significance: There are important common features in the networks involved in IEDs in patients with a heterogeneous range of epileptogenic foci. We confirm that the piriform cortex is a common node underlying IEDs in patients with focal epilepsy and so provides a target for further study and potential therapy. We describe important features of BOLD signal change that accompany focal and diffuse IEDs that will help researchers and clinicians navigate the sometimes complex findings revealed by these studies. Ó 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

1. Introduction EEG combined with fMRI (EEG–fMRI) is increasingly available to investigate interictal epileptiform discharges (IEDs) in patients with presumed focal epilepsy (Lazeyras et al., 2000; Diehl et al., 2003; Bagshaw et al., 2004; Aghakhani et al., 2006; Salek-Haddadi et al., 2006; Bonaventura et al., 2006; Zijlmans et al., 2007; De Tiege et al., 2007; Lui et al., 2008; Manganotti et al., 2008; Tyvaert et al., 2008; Jacobs et al., 2008; Moeller et al., 2009; LeVan et al.,

⇑ Corresponding author. Address: Brain Research Institute, PO Box 5444, Heidelberg Heights, 3081 Victoria, Australia. Tel.: +61 390357068. E-mail address: [email protected] (G.D. Jackson).

2010; Rathakrishnan et al., 2010; Thornton et al., 2010; Borelli et al., 2010; Pesaresi et al., 2011; Grouiller et al., 2011; Laufs et al., 2011; van Houdt et al. 2013) One reason this difficult to obtain data is important is the hope that results from these studies will provide an insight into the substrates underlying focal epilepsy thereby improving patient management (Zijlmans et al., 2007; Moeller et al., 2009; Thornton et al., 2010; Van Houdt et al., 2013). The focal epilepsies are a group of conditions that account for 60% of all epilepsies (Banerjee et al., 2009), and some of these patients can be rendered seizure free by resection of the epileptic focus. For this reason the aim of the extensive and detailed investigations that are routinely carried out on patients with focal

1388-2457/$36.00 Ó 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.clinph.2013.06.182

22

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

epilepsy (e.g. EEG, MRI, SPECT and PET) is to provide information that may help identify the epileptogenic zone. This zone is defined as ‘‘the minimum amount of cortex that must be resected (or completely disconnected) surgically to produce seizure freedom’’ (Luders et al., 2006) and so accurate identification of the anatomical limits of this zone is the dominant issue in the presurgical work-up of patients with focal epilepsy, but this has proven to be an elusive goal and there is no single ‘gold standard’ test to define this area. EEG–fMRI is a powerful new tool in this quest and is typically used to study IEDs. IEDs are ideally suited for fMRI studies because they are the most common marker of epilepsy and are abundant, easily recognized EEG events that elicit identifiable and reproducible BOLD signal changes (Benar et al., 2002; Gholipour et al., 2011; Pesaresi et al., 2011). Furthermore, they are typically sub-clinical, that is to say that the subject will not move because of one of these events and therefore the fMRI data will not be confounded by event related motion. While IEDs are ideally suited to EEG–fMRI studies, they may not delineate the epileptogenic zone (Marsh et al., 2010). IEDs arise from a region defined as the irritative zone (Luders et al., 2006) that may (or may not) overlap with the epileptogenic zone. IEDs may be more extensive, sometimes spreading into other cerebral lobes, and even into the contralateral hemisphere (Alarcon et al., 1997; Labate et al., 2004; Marsh et al., 2010; Vulliemoz et al., 2010). Furthermore a variety of IEDs may arise from adjacent normal cortex in the ‘irritative zone’. Understanding the reasons for the variety of IEDs and the pathways by which they spread across lobes and hemispheres will have consequences for our understanding of basic epileptic processes as well as direct implications when trying to use EEG–fMRI as a clinical tool. In the current study we review our experience acquired from all patients with focal epilepsy studied at high field (3T) in our centre over a seven year period paying particular attention to electrographic field of the IEDs and the common structures and therefore potential networks involved in the generation, spread and regulation of these events. We reviewed individual results and assessed their utility in the management of individual cases, hypothesising that the fMRI result would largely reflect the EEG discharge but provide better spatial information. We also conducted group analysis to examine the hypothesis that focal IEDs have a common underlying substrate.

having typical IEDs during the fMRI scan together with good quality MRI and EEG. Patients with an electro-clinical diagnosis of focal epilepsy who did not have any epileptiform discharges recorded inside the scanner or with unsatisfactory technical quality of the EEG recording are not included. Clinical details are summarized in Table 1. Clinical features were based on review of patient’s clinical notes. The study was approved by the ethics committee of our institution and all subjects gave informed consent, including parental consent from subjects under the age of 18 years. 2.2. EEG recording Eighteen non metallic scalp electrodes with carbon fibre leads were attached to the scalp in the conventional ‘10–20’ locations (with the exception of Fz). ECG was recorded from 1 or 2 electrodes placed on the chest. EEG data were acquired using a custom built amplifier with fibre optic transfer of data to a computer in the MR control room to allow on line monitoring of the EEG signal. MR gradient artifacts in the EEG signal was minimized by hardware design, and residual artifact removed off line, and in more recent studies, ballistocardiogram and movement artifacts were also reduced using head movement detection coils (Masterton et al., 2007). Real-time display, filtering and recording were performed using software developed in-house running on a Windows XP computer. The entire recording was conducted without activation procedures, and the patient was encouraged to sleep. EEG was recorded briefly outside the scanner to ensure good technical quality and identification of interictal discharges (IEDs) prior to continuous 3T EEG–fMRI acquisition for up to one hour. 2.3. MRI acquisitions fMRI data were obtained using a 3-T GE Signa LX whole body scanner (General Electric, Milwaukee, Wisconsin) with continuous acquisition of gradient-recalled echo planar image volumes (TR = 3200 ms, TE = 40 ms, flip angle = 80° with axial oblique slices 3.2 mm thick + 0.2 mm gap, 22 cm FOV; 64  64 matrix). The first 4 image acquisitions were discarded to ensure steady state tissue magnetization. fMRI data were acquired for 20–60 min (average 53 min). 2.4. Data analysis

2. Materials and methods 2.1. Patients In our centre, a total of 218 patients with a presumptive diagnosis of epilepsy (any type, including focal) were studied with EEG–fMRI at 3T between the years 2003–2010. From that data set we reviewed the data for all the patients with an electro-clinical diagnosis of focal epilepsy studied during that time (79 patients). The patients were mostly referred from the Comprehensive Epilepsy Programs at the Austin Hospital as well as some from the Royal Melbourne Children’s Hospital. These are tertiary referral venues for characterization of epilepsy syndromes and assessment of suitability of patients with refractory focal epilepsy for surgery. Diagnosis was based on ILAE criteria for focal epilepsy and was confirmed by reviewing the clinical, imaging and EEG data in detail for those patients. The basis of the current report is the findings in 33 patients with a confirmed diagnosis of focal epilepsy (16 females; mean age: 28 years, range 10–49 years) who had a successful EEG–fMRI study (the remaining 46 patients did not have IEDs during the monitoring period). A successful EEG–fMRI study was one defined as

2.4.1. EEG The whole EEG record was reviewed offline by an experienced electroencephalograher for identification of ictal and interictal epileptiform discharges (IEDs). IEDs were defined as events with the same field, duration and EEG waveform morphology as confirmed IEDs seen on routine EEG. Ambiguous EEG events which could not be confidently classified as epileptiform were identified and included in the subsequent analysis as ‘events of no interest’, but not considered further in the present study. It is important to note that while patients with a diagnosis of focal epilepsy often have focal IEDs, many will also have a variety of more diffuse IED types. We have classified IEDs based on their spatial distribution. Because of the relative limited EEG cover (10–20 electrode positions with the exception of Fz) we have limited the specificity of our EEG localization (i.e. we do not claim any localization beyond the lobar level). 2.4.2. EEG classification Focal discharges were defined as IEDs with a field that was consistent with an origin limited to a single lobe (i.e. they may be frontal, temporal, parietal or occipital). In this regard we also accepted

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

23

Table 1 Clinical details. Pt #

Age

Sex

Routine interictal EEG

Electro-clilical localization

Imaging findings

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

17 38 44 29 35 49 17 35 27 11 33 10 16 40 41 16 25 17 21 38 31 34 10 13 18 49 16

M M F M M F M F M M F F F F F F M F M F M M F F F M M

indep R and L front, cent R temp, diff PFA diff SW, R > L indep R front, L occip L occip, diff slow R front, diff SW R post L temp, L cent indep L and R temp diff SW R > L R temp R diff L cent, diff SW diff L > R diff R indep L and R front L cent diff cent diff L > R R post L temp diff front R > L L temp L cent diff R > L diff L SW diff R > L L temp SW

R par R temp diff R uncertain L post R front/temp multifocal epilepsy L cent L post diff L R post L post diff L R Front R cent L cent uncertain uncertain R temp L temp R front L temp diff L uncertain diff L R front/temp L par

MRI multi-R lesions (?FCD); PET concord; SPECT concord MRI mesial/lateral temp abn (?infarct); PET concord; SPECT concord MRI R hipp hypertrophy SPECT diff; PET NAD MRI multi bilat tubers MRI NAD; PET L temp; SPECT R post MRI R front and temp abn; PET concord; SPECT R temp MRI R front cavernoma; PET R temp; SPECT R front, L cingulate MRI NAD; PET NAD; SPECT NAD MRI NAD; PET L insula; SPECT, bil temp MRI NAD; PET, diff L; SPECT diff L > R MRI bil infarct, R > L; PET concord, SPECT NAD MRI NAD; PET L temp; SPECT L temp MRI NAD; PET L temp MRI R front MRI NAD MRI NAD; PET L cent MRI NAD; PET diff bilateral; SPECT NAD MRI NAD; SPECT L cent MRI NAD MRI L PVNH; PET bitemp L > R; SPECT L temp MRI NAD; PET NAD MRI NAD; PET L temp; SPECT L temp MRI Bi-perisylvian PMG Double cortex MRI NAD MRI NAD; PET bil; SPECT R front MRI NAD; PET L par; SPECT L par

Abn = abnormality; Bil = bilateral; Cent = central; Concord = concordant; Diff = diffuse; F = female; FCD = focal cortical dysplasia; Front = frontal; Hipp = hippocampal; Indep = independent; L = left; Temp = temporal; M = male; NAD = no abnormality detected; Occip = occipital; Par = parietal; PFA = paroxysmal fast activity; PMG = polymicrogyria; Post = posterior; Pt# = patient number; PVNH = peri-ventricular nodular hetereotopia; R = right; SW = spike and wave.

discharges that may appear to have a limited bilateral field if they arose from near the midline. Lateralized discharges were defined as IEDs that were restricted to a single hemisphere, but spread over more than one lobe. Diffuse bilateral discharges were defined as IEDs with a bilateral distribution that extended to involve one or more lobes in both hemispheres.

2.4.3. fMRI analysis fMRI data were pre-processed and analysed using SPM8 (Statistical Parametric Mapping, Welcome Department of Imaging Neuroscience, London, United Kingdom (http://www.fil.ion.ucl.ac.uk/ spm/software/spm8/). Standard pre-processing steps were: slice-timing correction; rigid-body realignment to correct for subject motion; and spatial smoothing with a Gaussian kernel (FWHM = 6 mm for individuals, 8 mm for the group analysis: Waites et al., 2005; Lemieux et al., 2008). The realignment parameters were incorporated in the analysis to model subject motion (Friston et al., 1996). The annotations on the study EEG were used to determine the onset time and duration of epileptiform activity in a ‘box-car’ event related analysis. IED duration was estimated based on the duration from the beginning to the end of the relevant waveform. For this reason we report ‘event duration’ and do not report ‘spike numbers’ so as to better accommodate the variation seen in typical IEDs (some very brief isolated spikes and some longer epileptiform complexes such as poly-spike, spike and wave, sharp and wave and paroxysmal fast of variable duration). The canonical hemodynamic response function (HRF) including temporal and dispersion derivatives was used to model event related BOLD signal changes. In order to identify robust regional increases and decreases in BOLD signal changes associated with epileptiform events, voxels were thresholded at p < 0.05 corrected for multiple comparisons and the resulting Statistical Parametric Map superimposed on an average echo planar image created during the course of image preprocessing.

2.4.4. BOLD signal classification Cortical BOLD signal changes were also classified according to the extent of activation/deactivation as focal, lateralized and diffuse bilateral, following similar rules to those applied to IEDs. The difference was that, unlike EEG where we are forced to infer localization from the IED field, when we consider significant BOLD signal change, we can more accurately identify anatomical location and therefore be more precise in our classification. Focal BOLD signal changes were defined as studies where all significant voxels arose within a single lobe. Lateralized BOLD signal changes were defined as studies where all significant voxels were restricted to a single hemisphere, but spread over more than one lobe. Diffuse bilateral discharges were defined as studies where the significant voxels had a bilateral distribution that extended to involve one or more lobes in both hemispheres. Two further clarifications were added to this classification in the form of ‘laterality emphasis’ (in cases where a bilateral diffuse pattern had a clear asymmetry) and identification of ‘multi-nodal patterns’. A node was arbitrarily defined as a discrete cluster of greater than 100 voxels with significant BOLD signal change within a more extensive pattern (see studies 14 and 21 in Fig. 1 for examples of lateralized BOLD patterns with bi-nodal patterns). BOLD signal changes in subcortical structures was examined in detail and dealt with separately. IED event durations and numbers of suprathreshold voxels (separated according to cortical vs subcortical distribution) were quantified and compared using t-tests. 2.4.5. Random effects group level fMRI analyses For group analyses we conducted random effects second level analyses. For these analyses we modified our preprocessing to suit this type of analysis by using 8 mm spatial smoothing and data normalized to an in-house calculated average template. Furthermore, in keeping with the methods used by Laufs et al. (2011) and Fahoum et al. (2012) in their second order analysis of patients with focal epilepsy, we manipulated the images in order to

24

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

Fig. 1. Representative examples of significant BOLD signal changes (p < 0.05, corrected for multiple comparisons) obtained from the event related analysis of individual categories of IEDs. The thresholded statistical parametric map is superimposed on the mean echo planar image obtained from each respective patient. The alphanumeric label on each Figure first presents the patient and next study being displayed (e.g. in the first image ‘3b’ refers to patient #3, study b). Next, beside the study identifier is a description of the IEDs analysed. Beside each fMRI/anatomical image is an example of the EEG obtained during the EEG–fMRI study obtained from each patient. The images are presented in radiological convention. R = right, L = left. The colour bar at bottom left represents the statistical scale. The axis at bottom right represents EEG calibration (10 lV vs 1 s). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

25

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

maintain consistent lateralization of the epileptic focus (we chose left as the preferred lateralization). To achieve this we identified those studies that had right sided IEDs and we flipped those images along the x-axis prior to normalization. In this way the side ipsilateral to the EEG abnormality was always on the left. A random effects model was then employed using the resulting t-contrasts to identify common regions across the group. 3. Results 3.1. IEDs studied The 33 patients gave a total of 51 IED datasets for analysis (12 patients had more than one type of epileptiform discharge). Of those 33 patients, 27 (82%) had statistically significant changes in BOLD signal associated with their 39 IED datasets. That is to say, statistical analysis of 12 IED datasets in six patients did not demonstrate any significant change in BOLD signal. Seven of those had ten seconds or less of IEDs compared to a mean of 110s in the cases with activations. In the remainder of this report we only consider the datasets from the 27 patients with significant (p < 0.05 corrected for multiple comparisons) BOLD signal change associated with IEDs.

The IED subtype events were: sharp-slow, spike-wave, polyspikes, paroxysmal fast activity, and intermittent rhythmic slowing. IEDs in patients with ‘focal’ epilepsy did not always have discrete, focal events on the EEG and three of our patients had focal as well as diffuse bilateral IEDs (patients #1, 5 and 6), one patient had focal as well as lateralized IEDs (patient #8) and one had focal, lateralized and diffuse bilateral IEDs (patient #9). The details of the type of IEDs and corresponding EEG–fMRI are summarized in Table 2. Patterns of BOLD signal change varied across the cohort, with each individual being distinctive (see examples in Fig. 1), however some common features were evident across the group.

3.2. Positive BOLD signal change was the predominant finding Positive BOLD signal change was the predominant finding (Fig. 1). This was noted across all IED types (31 of 39 studies) and the positive BOLD signal overlapped with the presumed location of the IED in most cases (94% of studies with positive BOLD changes), although in two studies the pattern of BOLD signal change did not overlap and was clearly discordant with expected findings (6%, Fig. 1, study 4b and 21).

Table 2 IED vs BOLD. St #

1a 1b 2a 2b 3a 3b 4a 4b 5a 5b 6a 6b 7a 7b 8a 8b 8c 9a 9b 9c 9d 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

EEG

BOLD pattern

EEG vs cortical + BOLD

IED

IED focality

Dur (s)

Cortical BOLD

Cortical + BOLD focality

Subcortical BOLD

Bil front R cent Diff fast Diff SW Diff SW Diff fast R front L occ Bil cent Diff slow R front Bil diff R post Occ fast L cent L diff L temp L diff Bil diff R temp Bil post Diff SW Diff fast Diff SW Diff SW R post Cent Cent fast Cent Diff SW R diff L temp Bil front L diff L cent Bil front L diff R temp L temp

Diff Focal Diff Diff Diff Diff Focal Focal Focal Diff Focal Diff Focal Focal Focal Lat Focal Lat Diff Focal Focal Diff Diff Diff Diff Focal Diff Focal Diff Diff Lat Focal Diff Lat Focal Diff Lat Focal Focal

44 2 20 233 539 94 63 29 84 123 48 20 186 8 202 150 86 88 51 11 24 251 52 493 15 57 4 11 46 125 73 12 60 88 169 19 283 117 109

+ >> + + >> + >> >> + +> + + + >> +> >> + + >> + = + + >> + + = + >> + >> + + + >> + = >> + + +> + + + >> +> + >> + + + + >>

Diff, R > L R focal Diff Diff Diff Diff, R >> L Diff Diff L lat multi-node Diff R lat multi-node Diff, R >> L R focal R focal R focal R focal R focal L lat multi-node Diff, L >> R R lat multi-node L lat multi-node Diff Diff Diff Diff, L > R R lat multi-node R lat multi-node L focal Diff multi-node Diff Diff R >> L L focal R focal L lat multi-node Diff, L >> R Nil L focal R focal L lat multi-node

Th(+), Ce(+) Nil Th(+),Ca(+), BS( ), Ce(+/ ) Th(+), BS( ), Ce(+) Th(+), Ce(+) Th(+),Ca(+), Ce(+/ ) Nil Th(+) Ce(+) Th(+),Ca( ), Ce(+/ ) Th(+), Ca(+),BG(+) Th(+), Ca(+) Nil Nil BG(+), Ce(+) Th( ),Ca( ),BG( ) Th( ) Ce( ) Ce( ) Nil Nil Th(+), BS( ), Ce(+) Th(+), Ce(+) Th(+/-),Ca( ), BS( ),Ce(+/ ) Th(+), Ce(+) Th( ) Nil Ce(+) Nil Nil Nil BG(+) Nil Nil BG( ) Nil Ce(+) Nil Nil

+ >> + >> +

Overlap Overlap Overlap Overlap Overlap Overlap Overlap Discord Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Overlap Discord Overlap Overlap Overlap Overlap Overlap Overlap

Bil = bilateral; Cent = central; Diff = diffuse; Dur = summed duration of all IEDs of each class recorded during EEG–fMRI study; FCD = focal cortical dysplasia; Front = frontal; L = left; Temp = temporal; Occip = occipital; Par = parietal; Post = posterior; R = right; SW = spike and wave; = negative BOLD signal; + = positive BOLD signal; > = greater than; >> = much greater than; BG = basal ganglia; BS = brainstem; Ca = caudate; Ce = cerebellum; Discord = discordant; Lat = lateralized; St# = study number (i.e. patient number plus study identifier); Th = thalamus.

26

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

3.3. The pattern of positive BOLD signal reflects the IED field Focal positive BOLD signal changes were predominantly associated with focal IEDs (of 11 studies with focal BOLD signal changes, eight arose from focal IEDs). By comparison, when negative BOLD signal changes were observed with focal IEDs, those negative BOLD signal changes were generally non localizing, being bilateral or absent in 15 studies and focal (and concordant) in only two studies. Lateralized IEDs, that is IEDs that were restricted to a single hemisphere but that spread over more than one lobe (studies 8a, 9a, 19, 22 and 25) were less likely to result in focal positive BOLD signal change, although in all cases there were ‘nodes’ of BOLD signal increase that overlapped with the presumed IED location. For example in study 9a (Fig. 1) the IED had the form of broad left hemisphere discharges, and the resulting analysis revealed five non-contiguous nodes of positive BOLD signal (two nodes of >100 voxels are apparent on this figure), and these were all located in the left hemispheres and so concordant with the broad lateralized IED field. Negative BOLD signal change was observed in all five

studies with lateralized IEDs and was lateralized and concordant in four of those studies. Diffuse IEDs also generally resulted in lateralized or diffuse patterns of positive BOLD signal (Table 2), although again some of these suggest a focal emphasis (e.g. Fig. 1, study 21). In studies with diffuse IEDs, we found that negative BOLD signal changes were generally non-localizing and bilateral (12 studies) or absent (three studies). In two studies the negative BOLD signal change was focal but discordant with other neuroclinical findings. 3.4. Positive BOLD signal in the thalamus was more common with diffuse IEDs The other notable finding was that positive BOLD signal within the thalamus was much more commonly seen in association with diffuse IEDs, (seen in 73% of diffuse IEDs). Of our 13 studies that showed positive thalamic BOLD signal, 11 were associated with diffuse IEDs and only 2 with focal IEDs (there were six other studies of diffuse IEDs that did not reveal thalamic activation). Significant BOLD signal change was most common in the mesial and

Fig. 2. Detailed images centred on mesial structures for all studies with significant changes in BOLD signal within the thalamus. Image labels and Figure legend are the same as described in Fig. 1.

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

27

Fig. 3. Detailed images centred on basal structures for all studies with significant changes in BOLD signal within the cerebellum. Image labels and Figure legend are the same as described in Fig. 1.

anterior thalamic regions (12 studies, Fig. 2), although precise localization was sometimes compromised because of the high signal associated with draining veins (the internal cerebral and thalamostriate veins). Our impression was that diffuse IEDs more often resulted in positive BOLD signal in mesial areas, whereas focal or lateralized discharges involved other thalamic regions. 3.5. Thalamic involvement was associated with co-involvement of other subcortical structures and was sometimes unilateral When thalamic involvement was observed (Table 2: 13 with positive and three with negative BOLD signal change) we often identified involvement of other subcortical structures. In 10 studies with thalamic involvement we also noted cerebellar involvement (Table 2: Fig. 3), in 7 there was caudate involvement, in 4 there was brainstem involvement and in 2 there was basal ganglia involvement. Subcortical involvement was also correlated with a higher number of suprathreshold voxels within the cortical mantle (p < 0.05). That is to say that when there was subcortical involvement, there were significantly more suprathreshold cortical voxels. Longer individual events were more likely to result in cerebellar involvement (mean event duration for studies with cerebellar involvement was 4.1s compared to those studies without cerebellar involvement 1.8 s p < 0.05 t test). In the fifteen studies where there was cerebellar involvement, only five did not have coinvolvement of the thalamus (Table 2). Cerebellar involvement was most often seen in the lateral part of the cerebellar cortex (Fig. 3).

In six patients there was evidence of a clear asymmetry in BOLD signal within some subcortical structures (thalamus, caudate and basal ganglia). When this was observed the laterality was always concordant with the electroclinical lateralization. It should be noted however that lateralised activation did not apply to all subcortical structures within an individual as sometimes there was bilateral activation in some structures and unilateral or lateralised activation in others. For example, in patient 8 with a presumed left sided electroclinical focus there was bilateral thalamic involvement, but unilateral involvement of the left caudate. 3.6. Within individual patients, diffuse IEDs had concordant, but more extensive BOLD signal increases when compared to focal IEDs Five patients had discrete focal as well as non-focal IEDs (pts 1, 5, 6, 8 and 9). The focal IEDs in these patients showed concordant areas of cortical positive BOLD signal (either focal or lateralized and overlapping with IED localization). On the other hand, the lateralized or diffuse bilateral IEDs in those patients showed more diffuse patterns of positive BOLD signal change, that overlapped with the focal results in cortical and subcortical components (Fig. 4). 3.7. Random effects group analysis confirms that the piriform cortex is commonly involved in focal discharges across the cohort Random effects group analysis of all studies (with those with right IEDs spatially flipped so that all apparent IEDs were ipsilat-

28

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

Fig. 4. Representative examples of the overlap in significant BOLD signal changes (p < 0.05, corrected for multiple comparisons) when the results of focal IEDs are compared to diffuse IEDs in the same patients. The legend in the bottom left panel represents the colours attributed to each condition.

area of the piriform, or pre-piriform cortex. A smaller area of significant positive BOLD signal change was also observed on the boundary of and also near the posterior horn of the lateral ventricle (inset in Fig. 5). Positive cortical BOLD signal often overlaps with identified structural and functional abnormalities. Furthermore in lesion negative cases areas of positive BOLD signal can suggest testable hypotheses as to the location of potentially occult lesions. Structural abnormalities were identified in 11 patients (Table 1) and in nine of those there were areas of positive BOLD signal changes that overlapped with, or were adjacent to the lesions in those patients. PET or SPECT studies identified functional abnormalities in 11 other patients who had no clear abnormality on MRI and in nine of those patients there were areas of positive BOLD signal changes that overlapped with or were adjacent to these functional abnormalities. In five other patients without clear abnormalities on imaging (MRI, PET and SPECT) we identified focal or lateralized areas of BOLD signal change on review of their EEG–fMRI studies.

Fig. 5. The result of the random effects group analysis (p < 0.05, corrected for multiple comparisons) superimposed on a mean image when all studies resulting from right sided IEDs were flipped so that the laterality of the data is homogeneous (i.e. all studies are oriented as if they had left sided IEDs). This reveals a clear increase in BOLD signal in the area of the ipsilateral piriform cortex. There are two other small areas of significant positive BOLD near the posterior margin of the lateral ventricles. There were no significant areas of negative BOLD signal.

eral) revealed a highly significant area of positive BOLD signal change (p < 0.05 FWE) in the ipsilateral frontal quadrant (Fig. 5). We reviewed these results in detail by visual inspection and used anatomical landmarks to determine localization. From this we concluded that the precise area of anterior signal change was in the

4. Discussion This report presents a review of our experience of the clinical and scientific utility of EEG–fMRI studies in patients with focal epilepsy using results from individual studies and from random effects group analysis. We confirm that EEG–fMRI of focal discharges have clinical relevance, showing focal findings in some patients. We also confirm the findings of Laufs et al. (2011) and Fahoum et al. (2012) that there are common nodes underlying IEDs in patients with focal epilepsy. Given the innate heterogeneity of the focal epilepsies, this is an unexpected and somewhat non-intuitive finding but our study reasserts these observations.

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

Like Laufs et al. (2011) we conducted group analysis on IEDs in a population of patients which had focal epilepsy but were heterogeneous in localisation and demonstrated a highly significant common node across the group. Review of our resulting images leads us to conclude that the area of positive BOLD signal is most likely to be within the ipsilateral piriform or pre-piriform cortex and this is in keeping with the localization revealed by Laufs although, like Laufs, none of our subjects had MRI abnormalities within this region. Furthermore, this cortical area was not consistently observed in the individual statistical parametric maps of our patients. This leads us to conclude that the piriform cortex is a common network node of spread and/or regulation that is common to focal epilepsies, rather than a generator of seizures itself. The piriform cortex (sometimes called the ‘area tempestas’) is an epileptogenically sensitive area in animal kindling models of temporal lobe epilepsy (Piredda and Gale, 1985; Racine et al., 1988; Löscher and Ebert, 1996) and recent work has also revealed that spread of limbic seizures into the piriform cortex is part of the seizure pathway for recruitment of the frontal lobe networks (McIntyre and Gilby, 2008). The identification of the piriform cortex as a common node in human IEDs therefore provides an important confirmation of the postulated networks revealed by the animal models of temporal lobe epilepsy. In both our data set and that of Laufs et al. (2011) there were a substantial proportion of patients with either frank temporal abnormalities on MRI, or at least evidence of temporal dysfunction as reflected by temporal IEDs [15/27 in the current investigation, Table 1, and 13/19 patients in Laufs study; online Supplemental Table-1a in Laufs et al. (2011)]. This raises the question of a bias in our results introduced by the preponderance of patients with temporal lobe dysfunction. In our current study we had insufficient data to isolate this variable, however Fahoum et al. (2012) addressed this using a larger cohort with sufficient numbers to allow subdivision and analysis of their data based on three separate cohorts (temporal, frontal and posterior epilepsies). In that study Fahoum et al. (2012) found that a network of nodes underlay the IEDs in each of their cohorts, with very similar networks underlying IEDs in patients with temporal lobe epilepsy and in patients with frontal lobe epilepsy. In both cohorts there were significant activations in the anterior cingulate and cerebellum as well as in the inferior frontal quadrant in the area of the frontal operculum and insula, furthermore in the cohort with frontal lobe epilepsy there was a significant activation within the medial thalamus. The highly significant node in the area of the ‘insula’ would be consistent with the activation reported by our study and the study of Laufs. The networks identified by Fahoum however, are more extensive than what we report and there are differences between the frontal and temporal cohorts, in particular with respect to cerebellar and thalamic involvement. This difference in the extent of the network may be due to the differences in the statistical thresholds chosen for data review as well as differences in the homogeneity of the cohorts studied. While we did not find cingulate, cerebellar or thalamic nodes in our group analysis, we did observe them on review of the data in our individual patients. Although these components of networks may not be common to the whole group, they are likely to be critical for understanding the epileptic network in a given individual. Some subgroups may activate particular nodes. For example, anterior cingulate involvement was seen in studies 3b, 9a and c and 16 (Fig. 1) and thalamic involvement was identified in 13 studies (Fig. 2) although we did not observe any clear trend in patients with frontal lobe epilepsy. In fact, in our experience, thalamic involvement was not consistently associated with frontal dysfunction and only two of our seven patients who had frontal aspects to their electroclinical localization demonstrated thalamic involvement. Instead, we found that subcortical involvement (and in par-

29

ticular involvement of the thalamus and cerebellum) was more closely associated with the extent of the IED field (p < 0.05; i.e. if the IED field was broad, then there was more likely to be significant subcortical increases in BOLD signal). This implies an active or at least a regulating role for subcortical structures in the spread of the cortical component of IEDs, even in cases of focal epilepsy. Thalamic (and subcortical) BOLD signal change is often reported as a feature of diffuse or ‘generalized’ IEDs regardless of the syndrome e.g. (Aghakhani et al., 2006; Siniatchkin et al., 2007; Carney et al., 2010) and it seems likely that these subcortical changes are better correlated with the spatial extent of the IED than with the underlying clinical syndrome. In our experience most discrete, focal IEDs give rise to focal cortical areas of BOLD signal increase that is concordant with the location of the IED and do not demonstrate positive BOLD signal in subcortical structures, but most diffuse IEDs do, regardless of the underlying pathophysiology. There have been suggestions that the localization within the thalamic nuclei may be important. Aghakhani et al. (2006) examined the role of the thalamus in bilateral synchrony and reported predominantly mesial and dorsal thalamic involvement during bilaterally synchronous spikes. We also found evidence of mesial involvement when we studied diffuse discharges in our cohort. Furthermore, Siniatchkin et al. (2007) reported centromedian and anterior thalamic involvement in their study of patients with Lennox Gastaut syndrome. Thalamic involvement is also well established in studies of genetic generalized epilepsies GGE: (Bagshaw et al., 2004; Hamandi et al., 2006; Moeller et al., 2008a,b, 2011; Tyvaert et al., 2009; Vaudano et al., 2009; Carney et al., 2010; Szaflarski et al., 2010) and it has been suggested that the anterior thalamic nucleus may play a role in the maintenance of epileptic discharges in those patients (Tyvaert, et al. 2009). The findings of our current study, along with those of Aghakhani et al.(2006) and Siniatchkin et al. (2011) suggest that in patients without a genetic generalized epilepsy it is the mesial thalamic structures that are involved when IEDs have a more diffuse field (although some caution is required as increased signal in mesial thalamostriate draining veins sometimes confound these findings). While thalamic involvement may represent facilitation or regulation of the IED, the cerebellar involvement observed during diffuse IEDs may have a different significance. We found that positive BOLD signal was most often observed in the lateral lobes of the cerebellum and that cerebellar involvement was significantly more likely with longer duration IED discharges (p < 0.05). Cerebellar involvement has been frequently reported (De Tiege et al., 2007; Yonghong Liu TY, 2008; Vaudano et al., 2009; Szaflarski et al., 2010; Moeller et al., 2011; Benuzzi et al., 2012), but seldom discussed in detail. The cerebellum has been implicated in seizure generation in both animal (Julien and Laxer, 1974; Hablitz and Wray, 1977) and human studies (Harvey et al., 1996; Mesiwala et al., 2002; Minkin et al., 2008), with accompanying IEDs that are presumably conducted to the cortex from a cerebellar generator, however none of our patients had any evidence of cerebellar epilepsy onset in their clinical semiology. We postulate therefore, that it is the rich interconnections from the frontal and limbic circuitry to the cerebellum (Snider and Maiti, 1976; Middleton and Strick, 2000; Norden and Blumenfeld, 2002) that provides the most likely explanation for what we see in EEG–fMRI studies. We suggest that the pattern of limbic involvement revealed in EEG–fMRI studies would inevitably result in secondary involvement of the cerebellum which is also supported by the finding that cerebellar involvement is more likely with prolonged IED duration. What role the cerebellum plays in initiation, termination or regulation of epileptic activity by being part of this network is currently unknown. Interestingly, Fahoum et al. (2012) revealed that the cerebellar activation varied according to IED localization and was ipsilateral in the analysis of the temporal lobe cohort and contralateral in

30

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31

the frontal lobe cohort. What may be useful in future studies using larger, homogeneous cohorts would be to examine the distribution of cerebellar activation associated with different cortical and subcortical activation patterns. The common regions of ‘activation’ reported above are important to our broad understanding of the critical networks involved with IEDs in focal epilepsy, however we also made observations that were specific to individual patients and this leads us to discuss the clinical utility of EEG–fMRI in individual studies of focal epilepsy. Fundamentally it is important to note that EEG/fMRI studies reveal an aspect of focal epilepsy that is not available to other standard investigations (MRI, PET, SPECT and EEG) by identifying the 3-dimensional pattern of cortical and subcortical involvement associated with IED events. As such this method provides a useful, independent localization tool. In general we found (as have others) that areas of positive BOLD signal were more ‘useful’ (often focal and concordant) than areas of negative BOLD signal (which we found to be mostly absent or small and bilateral, although sometimes concordant and lateralized in lateralized IEDs; others have also suggested that negative BOLD signal changes were unhelpful, see Aghakhani et al., 2006; Yonghong Liu TY, 2008 and Siniatchkin et al., 2011). In addition, while we found that the results of EEG/ fMRI studies were often in agreement with other clinical investigations, there were times when they revealed important new localization information (in five of our studies the EEG/fMRI results implicated discrete cortical ‘targets’ in patients who had no lesion visible on standard imaging). Finally there were times when the results did not identify an expected solitary focus, but were either discordant or more complex, suggesting a network of cortex involved in the epileptic events. Results that may at first appear discordant (e.g. Fig. 1 pt #21 had parietal BOLD foci associated with bilateral frontal IEDs) can provide important new information and lead to alternative hypotheses regarding localization that in turn may affect management decisions. Similarly, hypotheses that are not restricted to a solitary focus may be required to explain studies where the results revealed complex patterns of BOLD signal change. For example, in some studies we identified distant cortical multi-nodal areas of BOLD signal change (e.g. Fig. 1 pt #14). While we have discussed above the role of subcortical structures in the spread and regulation of diffuse IEDs, multi-nodal cortical involvement often requires different hypotheses that may involve the concept of excitable cortico-cortical networks or the recruitment of inherent brain systems. Such multi-nodal cortical patterns may represent ‘benign’ propagation (involving inherent pre-existing networks such as the ‘REST’ networks; De Luca et al., 2006; Jann et al., 2010) or multifocal epileptic nodes, perhaps incorporating neuronal populations with electrophysiological involvement that is too deep or asynchronous to contribute to scalp EEG. Indeed, multi-nodal patterns of BOLD signal change have previously been reported in studies of IEDs in patients with focal epilepsy and such findings were not unexpected and as Gotman and Pittau (2011) noted, ‘‘(the BOLD signal). . . increase is most intense in the region generating the discharge but is also present in regions affected by the discharge’’. The identification multi-nodal BOLD signal changes can be clinically important by providing data for these alternative hypotheses. In one of our early studies (pt #14), resective surgery was undertaken based upon a structural frontal abnormality identified on MRI. Subsequent review of our EEG–fMRI results revealed that the MRI abnormality overlapped with an area of BOLD signal change but that there was also another significant node located in the ipsilateral parietal lobe. In this patient resective surgery did not improve seizure frequency and our current understanding of the EEG–fMRI data would suggest a cortico-cortical network that should be further investigated with methods such as intracranial

studies. The problem in cases such as these is obviously in differentiating the ‘generator’ from the areas of spread. In a recent study LeVan et al. (2010) looked at EEG amplitude modulation of BOLD responses and they found a direct correlation in the area of the ‘epileptogenic focus’ but no correlation at nodes distant from this. This suggests that further work may allow us to disentangle these complex patterns of BOLD signal change. In summary, our current investigation has revealed a powerful and novel role for EEG–fMRI studies as a clinical tool in focal epilepsy. The results must be viewed cautiously within the context of networks that may either underlay, or be affected by the spread of focal IEDs. At a group level we found that the piriform cortex is an underlying common node for focal IEDs and as such warrants further investigation that may be a common target for focal epilepsy therapies (Laufs et al., 2011), especially in cases where it is the IEDs themselves that represent the significant clinical problem (such as in chronic encephalopathies). Secondly, when faced with individual patient data, the localization revealed by the results of EEG–fMRI studies provides important, independent cortical and subcortical localization that is not apparent in raw scalp EEG. The localization identified from the BOLD signal may confirm findings from other studies (MRI, PET and SPECT), or may lead to plausible alternative explanations. The observation that some IEDs result in discrete (sometimes distant) multi-nodal areas of BOLD signal change either in the cortex or in subcortical structures does not imply that the patient is not a surgical candidate and most likely reflects the networks that accompany diffuse electrographic events even when driven by focal pathology. Such multi-nodal patterns of BOLD signal change may represent ‘benign’ propagation or multifocal epileptic nodes and some effort may be required to disentangle these alternatives. We found that diffuse or hemispheric IEDs in a patient with focal epilepsy will generally yield bilateral or lateralized patterns of BOLD signal change with thalamic and cerebellar involvement, but often within those patterns will be nodes of cortical ‘activation’ that provide important information that can aid in the clinical interpretation of the epileptogenic process and can guide the next steps of investigation or treatment. Conflict of interest The authors report no conflict of interest. Acknowledgements We gratefully acknowledge Shawna Farquharson. Renee Mineo, Heather Ducie, Nonie Morrish, Saba Ansari, Janet Barchett and Selma Music for assistance in acquiring and collecting these data as well as the epileptologists at Austin Health and the Royal Children’s Hospital Melbourne for referral of patients. This study was supported by the National Health and Medical Research Council of Australia (NHMRC Project Grant 628725) and the Operational Infrastructure Support Program of the State Government of Victoria. References Aghakhani Y, Kobayashi E, Bagshaw AP, Hawco C, Benar CG, Dubeau F, et al. Cortical and thalamic fMRI responses in partial epilepsy with focal and bilateral synchronous spikes. Clin Neurophysiol 2006;117:177–91. Alarcon G, Garcia Seoane JJ, Binnie CD, Martin Miguel MC, Juler J, Polkey CE, et al. Origin and propagation of interictal discharges in the acute electrocorticogram. Implications for pathophysiology and surgical treatment of temporal lobe epilepsy. Brain 1997;120:2259–82. Bagshaw AP, Aghakhani Y, Benar CG, Kobayashi E, Hawco C, Dubeau F, et al. EEG– fMRI of focal epileptic spikes: analysis with multiple haemodynamic functions and comparison with gadolinium-enhanced MR angiograms. Hum Brain Mapp 2004;22:179–92.

D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31 Banerjee PN, Filippi D, Allen Hauser W. The descriptive epidemiology of epilepsy-a review. Epilepsy Res 2009;85:31–45. Benar CG, Gross DW, Wang Y, Petre V, Pike B, Dubeau F, et al. The BOLD response to interictal epileptiform discharges. Neuroimage 2002;17:1182–92. Benuzzi F, Mirandola L, Pugnaghi M, Farinelli V, Tassinari CA, Capovilla G, et al. Increased cortical BOLD signal anticipates generalized spike and wave discharges in adolescents and adults with idiopathic generalized epilepsies. Epilepsia 2012;53:622–30. Bonaventura CD, Vaudano AE, Carni M, Pantano P, Nucciarelli V, Garreffa G, et al. EEG/fMRI study of ictal and interictal epileptic activity: methodological issues and future perspectives in clinical practice. Epilepsia 2006;47(Suppl. 5):52–8. Borelli P, Avesani M, Formaggio E, Storti SF, Zanoni T, Moretto G, et al. EEG–fMRI as a useful tool to detect epileptic foci associated with secondary bilateral synchrony. Seizure 2010;19:605–8. Carney PW, Masterton RA, Harvey AS, Scheffer IE, Berkovic SF, Jackson GD. The core network in absence epilepsy. Differences in cortical and thalamic BOLD response. Neurology 2010;75:904–11. De Luca M, Beckmann CF, De Stefano N, Matthews PM, Smith SM. FMRI resting state networks define distinct modes of long-distance interactions in the human brain. Neuroimage 2006;29:1359–67. Diehl B, Salek-haddadi A, Fish DR, Lemieux L. Mapping of spikes, slow waves, and motor tasks in a patient with malformation of cortical development using simultaneous EEG and fMRI. J Magn Reson Imaging 2003;21:1167–73. Fahoum F, Lopes R, Pittau F, Dubeau F, Gotman J. Widespread epileptic networks in focal epilepsies–EEG–fMRI study. Epilepsia 2012;53:1618–27. Friston KJ, Williams S, Howard R, Frackowiak RS, Turner R. Movement-related effects in fMRI time-series. Magn Reson Med 1996;35:346–55. Gholipour T, Moeller F, Pittau F, Dubeau F, Gotman J. Reproducibility of interictal EEG–fMRI results in patients with epilepsy. Epilepsia 2011;52:433–42. Gotman J, Pittau F. Combining EEG and fMRI in the study of epileptic discharges. Epilepsia 2011;52(Suppl. 4):38–42. Grouiller F, Thornton RC, Groening K, Spinelli L, Duncan JS, Schaller K, et al. With or without spikes: localization of focal epileptic activity by simultaneous electroencephalography and functional magnetic resonance imaging. Brain 2011;134:2867–86. Hablitz JJ, Wray DV. Cerebellar unit activity during generalized penicillin epilepsy in the awake cat. Exp Neurol 1977;56:189–99. Hamandi K, Salek-Haddadi A, Laufs H, Liston A, Friston K, Fish DR, et al. EEG–fMRI of idiopathic and secondarily generalized epilepsies. Neuroimage 2006;31:1700–10. Harvey AS, Jayaka P, Duchowny M, Resnick T, Prats A, Altman N, et al. Hemifacial seizures and cerebellar ganglioglioma: an epilepsy syndrome of infancy with seizures of cerebellar origin. Ann Neurol 1996;40:91–8. Jacobs J, Rohr A, Moeller F, Boor R, Kobayashi E, LeVan Meng P, et al. Evaluation of epileptogenic networks in children with tuberous sclerosis complex using EEG– fMRI. Epilepsia 2008;49:816–25. Jann K, Kottlow M, Dierks T, Boesch C, Koenig T. Topographic electrophysiological signatures of FMRI resting state networks. PLoS 2010;5:e12945. Julien RM, Laxer KD. Cerebellar responses to penicillin-induced cerebral cortical epileptiform discharge. Electroencephalogr Clin Neurophysiol 1974;37:123–32. Labate A, Briellmann RS, Harvey AS, Berkovic SF, Federico P, Kalnins RM, et al. Temporal lobe dysembryoplastic neuroepithelial tumour: significance of discordant interictal spikes. Epileptic Disord 2004;6:107–14. Laufs H, Richardson MP, Salek-Haddadi A, Vollmar C, Duncan JS, Gale K, et al. Converging PET and fMRI evidence for a common area involved in human focal epilepsies. Neurology 2011;77:904–10. Lazeyras F, Blanke O, Perrig S, Zimine I, Golay X, Delavelle J, et al. EEG-triggered functional MRI in patients with pharmacoresistant epilepsy. J Magn Reson Imaging 2000;12:177–85. Lemieux L, Laufs H, Carmichael D, Suresh Paul J, Walker MC, Duncan JS. Noncanonical spike-related BOLD responses in focal epilepsy. Hum Brain Mapp 2008;29:329–45. LeVan P, Tyvaert L, Gotman J. Modulation by EEG features of BOLD responses to interictal epileptiform discharges. Neuroimage 2010;50:15–26. Löscher W, Ebert U. The role of the piriform cortex in kindling. Prog Neurobiol 1996;50:427–81. Luders HO, Najm I, Nair D, Widdess-Walsh P, Bingman W. The epileptogenic zone: general principles. Epileptic Disord 2006;8(Suppl. 2):S1–9. Lui S, Ouyang L, Chen Q, Huang X, Tang H, Chen H, et al. Differential interictal activity of the precuneus/posterior cingulate cortex revealed by resting state functional MRI at 3T in generalized vs partial seizure. J Magn Reson Imaging 2008;27:1214–20. Manganotti P, Formaggio E, Milanese F, Gasparini A, Cerini R, Bongiovanni LG, et al. Continuous EEG–fMRI in patients with partial epilepsy and focal interictal slowwave discharges on EEG. J Magn Reson Imaging 2008;26:1089–100. Marsh ED, Peltzer B, Brown MW, Wusthoff C, Storm PB, Litt B, et al. Interictal EEG spikes identify the region of electrographic seizure onset in some, but not all, pediatric epilepsy patients. Epilepsia 2010;51:592–601. Masterton RAJ, Abbott DF, Fleming SW, Jackson GD. Measurement and reduction of motion and ballistocardiogram artefacts from simultaneous EEG and fMRI recordings. Neuroimage 2007;37:202–11.

31

McIntyre DC, Gilby KL. Mapping seizure pathways in the temporal lobe. Epilepsia 2008;49(Suppl. 3):23–30. Mesiwala AH, Kuratani JD, Avellino AM, Roberts TS, Sotero MA, Ellenbogen RG. Focal motor seizures with secondary generalization arising in the cerebellum. J Neurosurg 2002;97:190–6. Middleton FA, Strick PL. Basal ganglia and cerebellar loops: motor and cognitive circuits. Brain Res Rev 2000;31:236–50. Minkin K, Tzekov C, Naydenov E, Ivanov I, Kulev O, Romansky K, et al. Cerebellar gangliocytoma presenting with hemifacial spasms: clinical report, literature review and possible mechanisms. Acta Neurochir 2008;150:719–24. Moeller F, Maneshi M, Pittau F, Gholipour T, Bellec P, Dubeau F, et al. Functional connectivity in patients with idiopathic generalized epilepsy. Epilepsia 2011;52:515–22. Moeller F, Siebner HR, Wolff S, Muhle H, Boor R, Granert O, et al. Changes in activity of striato-thalamo-cortical network precede generalized spike wave discharges. Neuroimage 2008a;39:1839–49. Moeller F, Siebner HR, Wolff S, Muhle H, Granert O, Jansen O, et al. Simultaneous EEG–fMRI in drug-naive children with newly diagnosed absence epilepsy. Epilepsia 2008b;49:1510–9. Moeller F, Tyvaert L, Nguyen DK, Levan P, Bouthillier A, Kobayashi E, et al. EEG– fMRI: adding to standard evaluations of patients with nonlesional frontal lobe epilepsy. Neurology 2009;73:2023–30. Norden AD, Blumenfeld H. The role of subcortical structures in human epilepsy. Epilepsy Behav 2002;3:219–31. Pesaresi I, Cosottini M, Belmonte G, Maritato P, Mascalchi M, Puglioli M, et al. Reproducibility of BOLD localization of interictal activity in patients with focal epilepsy: intrasession and intersession comparisons. Magn Res Mater Phys 2011;24:284–96. Piredda S, Gale K. A crucial epileptogenic site in the deep prepiriform cortex. Nature 1985;317:623–5. Racine RJ, Mosher M, Kairiss EW. The role of the pyriform cortex in the generation of interictal spikes in the kindled preparation. Brain Res 1988;454:251–63. Rathakrishnan R, Moeller F, Levan P, Dubeau F, Gotman J. BOLD signal changes preceding negative responses in EEG–fMRI in patients with focal epilepsy. Epilepsia 2010;51:1837–45. Salek-Haddadi A, Diehl B, Hamandi K, Merschhemke M, Liston A, Friston K, et al. Hemodynamic correlates of epileptiform discharges: an EEG–fMRI study of 63 patients with focal epilepsy. Brain Res 2006;1088:148–66. Siniatchkin M, van Baalen A, Jacobs J, Moeller F, Moehring J, Boor R, et al. Different neuronal networks are associated with spikes and slow activity in hypsarrhythmia. Epilepsia 2007;48:2312–21. Siniatchkin M, Coropceanu D, Moeller F, Boor R, Stephani U. EEG–fMRI reveals activation of brainstem and thalamus in patients with Lennox–Gastaut syndrome. Epilepsia 2011;52:766–74. Snider RS, Maiti A. Cerebellar contributions to the Papez circuit. J Neurosci Res 1976;2:133–46. Szaflarski JP, Difrancesco M, Hirschauer T, Banks C, Privitera MD, Gotman J, et al. Cortical and subcortical contributions to absence seizure onset examined with EEG/fMRI. Epilepsy Behav 2010;18:404–13. Thornton R, Laufs H, Rodionov R, Cannadathu S, Carmichael DW, Vulliemoz S, et al. EEG correlated functional MRI and postoperative outcome in focal epilepsy. J Neurol Neurosurg Psychiatry 2010;81:922–7. De Tiege X, Laufs H, Boyd SG, Harkness W, Allen PJ, Clark CA, et al. EEG–fMRI in children with pharmacoresistant focal epilepsy. Epilepsia 2007;48:385–9. Tyvaert L, Chassagnon S, Sadikot A, Levan P, Dubeau F, Gotman J. Thalamic nuclei activity in idiopathic generalized epilepsy: An EEG–fMRI study. Neurology 2009;73:2018–22. Tyvaert L, Hawco C, Kobayashi E, LeVan P, Dubeau F, Gotman J. Different structures involved during ictal and interictal epileptic activity in malformations of cortical development: an EEG–fMRI study. Brain 2008;131:2042–60. van Houdt PJ, de Munck JC, Leijten FS, Huiskamp GJ, Colon AJ, Boon PA, et al. EEG– fMRI correlation patterns in the presurgical evaluation of focal epilepsy: a comparison with electrocorticographic data and surgical outcome measures. NeuroImage 2013;26. http://dx.doi.org/10.1016/j.neuroimage.2013.02.033. in press. Vaudano AE, Laufs H, Kiebel SJ, Carmichael DW, Hamandi K, Guye M, et al. Causal hierarchy within the thalamo-cortical network in spike and wave discharges. PLoS ONE 2009;4:e6475. Vulliemoz S, Carmichael DW, Rosenkranz K, Diehl B, Rodionov R, Walker MC, et al. Simultaneous intracranial EEG and fMRI of interictal epileptic discharges in humans. Neuroimage 2010;54:182–90. Waites AB, Shaw ME, Briellmann RS, Labate A, Abbott DF, Jackson GD. How reliable are fMRI–EEG studies of epilepsy? a nonparametric approach to analysis validation and optimization. Neuroimage 2005;24:192–9. Yonghong Liu TY. EEG–fMRI study of the ictal and interictal epileptic activity in patients with eyelid myoclonia with absences. Epilepsia 2008;49:2078–86. Zijlmans M, Huiskamp G, Hersevoort M, Seppenwoolde J-H, van Huffelen AC, Leijten FSS. EEG–fMRI in the preoperative work-up for epilepsy surgery. Brain 2007;130:2342–53.