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
b
Department of Neurology Austin Health, Heidelberg, Australia
c
Epilepsy Research Centre, Department of Medicine, University of Melbourne, Melbourne, Australia

a r t i c l e i n f o h i g h l i g h t s

Article history:
 EEG/fMRI is an important new tool for studying patients with focal epilepsy and the electrographic
Accepted 27 June 2013
Available online 17 July 2013
field of the interictal epileptiform discharges (IEDs) is reflected in the fMRI results with focal and dif-
fuse IEDs providing novel and important localization information.
 Amongst patients with a heterogeneous array of focal epilepsies, the piriform cortex is a common
Keywords:
EEG–fMRI node in the underlying networks associated with focal IEDs.
IED  In cases of diffuse IEDs we noted involvement of subcortical structures, in particular the thalamus and
Focal epilepsy cerebellum.
Piriform cortex

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 signifi-
cant 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 under-
lying 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 2010; Rathakrishnan et al., 2010; Thornton et al., 2010; Borelli

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, Heidel-
berg Heights, 3081 Victoria, Australia. Tel.: +61 390357068.
E-mail address: gjackson@brain.org.au (G.D. Jackson).
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 inves-
tigations 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 anatomi-
cal 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 reproduc-
ible 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 nor-
mal 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 understand-
ing 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 electro-
graphic field of the IEDs and the common structures and therefore
potential networks involved in the generation, spread and regula-
tion 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 anal-
ysis to examine the hypothesis that focal IEDs have a common
underlying substrate.
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 Epi-
lepsy 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 monitor-
ing period). A successful EEG–fMRI study was one defined as
having typical IEDs during the fMRI scan together with good qual-
ity 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 institu-
tion 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 hard-
ware 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 per-
formed using software developed in-house running on a Windows
XP computer. The entire recording was conducted without activa-
tion 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
2.4.1. EEG
The whole EEG record was reviewed offline by an experienced
electroencephalograher for identification of ictal and interictal epi-
leptiform 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 lim-
ited 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 con-
sistent with an origin limited to a single lobe (i.e. they may be fron-
tal, 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 17 M indep R and L front, cent R par MRI multi-R lesions (?FCD); PET concord; SPECT concord
2 38 M R temp, diff PFA R temp MRI mesial/lateral temp abn (?infarct); PET concord; SPECT concord
3 44 F diff SW, R > L diff R MRI R hipp hypertrophy SPECT diff; PET NAD
4 29 M indep R front, L occip uncertain MRI multi bilat tubers
5 35 M L occip, diff slow L post MRI NAD; PET L temp; SPECT R post
6 49 F R front, diff SW R front/temp MRI R front and temp abn; PET concord; SPECT R temp
7 17 M R post multifocal epilepsy MRI R front cavernoma; PET R temp; SPECT R front, L cingulate
8 35 F L temp, L cent L cent MRI NAD; PET NAD; SPECT NAD
9 27 M indep L and R temp L post MRI NAD; PET L insula; SPECT, bil temp
10 11 M diff SW R > L diff L MRI NAD; PET, diff L; SPECT diff L > R
11 33 F R temp R diff R post MRI bil infarct, R > L; PET concord, SPECT NAD
12 10 F L cent, diff SW L post MRI NAD; PET L temp; SPECT L temp
13 16 F diff L > R diff L MRI NAD; PET L temp
14 40 F diff R R Front MRI R front
15 41 F indep L and R front R cent MRI NAD
16 16 F L cent L cent MRI NAD; PET L cent
17 25 M diff cent uncertain MRI NAD; PET diff bilateral; SPECT NAD
18 17 F diff L > R uncertain MRI NAD; SPECT L cent
19 21 M R post R temp MRI NAD
20 38 F L temp L temp MRI L PVNH; PET bitemp L > R; SPECT L temp
21 31 M diff front R > L R front MRI NAD; PET NAD
22 34 M L temp L temp MRI NAD; PET L temp; SPECT L temp
23 10 F L cent diff L MRI Bi-perisylvian PMG
24 13 F diff R > L uncertain Double cortex
25 18 F diff L SW diff L MRI NAD
26 49 M diff R > L R front/temp MRI NAD; PET bil; SPECT R front
27 16 M L temp SW L par 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 = poly-
microgyria; 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 (Statis-
tical Parametric Mapping, Welcome Department of Imaging Neuro-
science, 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 sub-
ject 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 param-
eters 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 re-
port ‘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 tempo-
ral 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 superim-
posed 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 dif-
fuse 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 signif-
icant BOLD signal change within a more extensive pattern (see
studies 14 and 21 in Fig. 1 for examples of lateralized BOLD pat-
terns 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. Further-
more, 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.)
 D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31 25

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 ipsilat-
eral to the EEG abnormality was always on the left. A random ef-
fects 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 demon-
strate 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, poly-
spikes, paroxysmal fast activity, and intermittent rhythmic slow-
ing. 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 loca-
tion 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 # EEG BOLD pattern EEG vs cortical + BOLD

IED IED focality Dur (s) Cortical BOLD Cortical + BOLD focality Subcortical BOLD
1a Bil front Diff 44 + >> Diff, R > L Th(+), Ce(+) Overlap
1b R cent Focal 2 + R focal Nil Overlap
2a Diff fast Diff 20 + >> Diff Th(+),Ca(+), BS( ), Ce(+/ ) Overlap
2b Diff SW Diff 233 + >> Diff Th(+), BS( ), Ce(+) Overlap
3a Diff SW Diff 539 >> + Diff Th(+), Ce(+) Overlap
3b Diff fast Diff 94 +> Diff, R >> L Th(+),Ca(+), Ce(+/ ) Overlap
4a R front Focal 63 + Diff Nil Overlap
4b L occ Focal 29 + Diff Th(+) Discord
5a Bil cent Focal 84 + >> L lat multi-node Ce(+) Overlap
5b Diff slow Diff 123 +> Diff Th(+),Ca( ), Ce(+/ ) Overlap
6a R front Focal 48 >> + R lat multi-node Th(+), Ca(+),BG(+) Overlap
6b Bil diff Diff 20 + >> Diff, R >> L Th(+), Ca(+) Overlap
7a R post Focal 186 + = R focal Nil Overlap
7b Occ fast Focal 8 + R focal Nil Overlap
8a L cent Focal 202 + R focal BG(+), Ce(+) Overlap
8b L diff Lat 150 >> + R focal Th( ),Ca( ),BG( ) Overlap
8c L temp Focal 86 + = R focal Th( ) Overlap
9a L diff Lat 88 + >> L lat multi-node Ce( ) Overlap
9b Bil diff Diff 51 + >> Diff, L >> R Ce( ) Overlap
9c R temp Focal 11 + R lat multi-node Nil Overlap
9d Bil post Focal 24 + L lat multi-node Nil Overlap
10 Diff SW Diff 251 + >> Diff Th(+), BS( ), Ce(+) Overlap
11 Diff fast Diff 52 + = Diff Th(+), Ce(+) Overlap
12 Diff SW Diff 493 >> + Diff Th(+/-),Ca( ), BS( ),Ce(+/ ) Overlap
13 Diff SW Diff 15 + Diff, L > R Th(+), Ce(+) Overlap
14 R post Focal 57 +> R lat multi-node Th( ) Overlap
15 Cent Diff 4 + R lat multi-node Nil Overlap
16 Cent fast Focal 11 + L focal Ce(+) Overlap
17 Cent Diff 46 + >> Diff multi-node Nil Overlap
18 Diff SW Diff 125 +> Diff Nil Overlap
19 R diff Lat 73 + >> Diff R >> L Nil Overlap
20 L temp Focal 12 + L focal BG(+) Overlap
21 Bil front Diff 60 + R focal Nil Discord
22 L diff Lat 88 + L lat multi-node Nil Overlap
23 L cent Focal 169 + >> Diff, L >> R BG( ) Overlap
24 Bil front Diff 19 Nil Nil Overlap
25 L diff Lat 283 + >> L focal Ce(+) Overlap
26 R temp Focal 117 + >> R focal Nil Overlap
27 L temp Focal 109 + L lat multi-node Nil 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 associ-
ated 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 sig-
nal 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 lateral-
ized IED field. Negative BOLD signal change was observed in all five
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.
studies with lateralized IEDs and was lateralized and concordant in
four of those studies.
Diffuse IEDs also generally resulted in lateralized or diffuse pat-
terns 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 stud-
ies of diffuse IEDs that did not reveal thalamic activation). Signifi-
cant BOLD signal change was most common in the mesial and
 D. Flanagan et al. / Clinical Neurophysiology 125 (2014) 21–31
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 sig-
nal associated with draining veins (the internal cerebral and thal-
amostriate veins). Our impression was that diffuse IEDs more
often resulted in positive BOLD signal in mesial areas, whereas fo-
cal 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 stud-
ies with thalamic involvement we also noted cerebellar involve-
ment (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 involve-
ment, 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 cerebel-
lar involvement 1.8 s p < 0.05 t test). In the fifteen studies where
there was cerebellar involvement, only five did not have co-
involvement of the thalamus (Table 2). Cerebellar involvement
was most often seen in the lateral part of the cerebellar cortex
(Fig. 3).
27
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 sub-
cortical 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 involve-
ment, 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.
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 con-
cluded that the precise area of anterior signal change was in the
area of the piriform, or pre-piriform cortex. A smaller area of signif-
icant positive BOLD signal change was also observed on the bound-
ary 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 neg-
ative 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 abnor-
malities 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.
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 ef-
fects 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-intu-
itive 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 heteroge-
neous in localisation and demonstrated a highly significant com-
mon 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 re-
gion. 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 epilep-
sies, 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 al-
low 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 cerebel-
lum 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 ‘in-
sula’ would be consistent with the activation reported by our study
and the study of Laufs. The networks identified by Fahoum how-
ever, are more extensive than what we report and there are differ-
ences 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 sta-
tistical 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 crit-
ical for understanding the epileptic network in a given individual.
Some subgroups may activate particular nodes. For example, ante-
rior 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 dysfunc-
tion and only two of our seven patients who had frontal aspects to
their electroclinical localization demonstrated thalamic involve-
ment. 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 syn-
drome 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, fo-
cal 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 syn-
chronous 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 tha-
lamic 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 epi-
lepsy 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 some-
times confound these findings).
While thalamic involvement may represent facilitation or regu-
lation of the IED, the cerebellar involvement observed during dif-
fuse 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 signifi-
cantly 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; Szaflar-
ski et al., 2010; Moeller et al., 2011; Benuzzi et al., 2012), but sel-
dom 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 epi-
lepsy 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 cerebel-
lum which is also supported by the finding that cerebellar involve-
ment is more likely with prolonged IED duration. What role the
cerebellum plays in initiation, termination or regulation of epilep-
tic 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 sub-
cortical activation patterns.
The common regions of ‘activation’ reported above are impor-
tant 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 epi-
lepsy. 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 some-
times 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 investiga-
tions, there were times when they revealed important new locali-
zation 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 re-
sults did not identify an expected solitary focus, but were either
discordant or more complex, suggesting a network of cortex in-
volved 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 manage-
ment 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 regula-
tion 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 neuro-
nal populations with electrophysiological involvement that is too
deep or asynchronous to contribute to scalp EEG. Indeed, multi-no-
dal 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 clin-
ically important by providing data for these alternative hypothe-
ses. 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 differen-
tiating 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 epi-
lepsy. 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 epi-
lepsy 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 cor-
tex 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 Sel-
ma Music for assistance in acquiring and collecting these data as
well as the epileptologists at Austin Health and the Royal Chil-
dren’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 Gov-
ernment 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 slow-
wave 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.