Understanding the pathophysiology of reflex epilepsy using simultaneous EEG-fMRI

Journal Identification = EPD Article Identification = 0632 Date: March 25, 2014 Time: 4:25 pm

do

i:10.

1684

/ep

d.2

014.

0632

Epileptic Disord, Vol. 16, No. 1, March 2014

Correspondence:Rose Dawn BharathCognitive Neuroscience Center and Deptof Neuroimaging and InterventionalRadiology,National Institute of Mental Healthand Neuro Science (NIMHANS),Hosur Road,Bangalore 560029, India<[email protected]>

Original articleEpileptic Disord 2014; 16 (1): 19-29

Understandingthe pathophysiologyof reflex epilepsyusing simultaneous EEG-fMRI

Manglore Sandhya 1,2, Rose Dawn Bharath 1,2,Rajanikant Panda 1,2, S.R. Chandra 3, Naveen Kumar 2,Lija George 1,2, A. Thamodharan 1,2, Arun Kumar Gupta 1,2,P. Satishchandra 3

1 Cognitive Neuroscience Center, National Institute of Mental Health, Neuro Science(NIMHANS)2 Dept of Neuroimaging and Interventional Radiology, National Institute of MentalHealth and Neuro Science (NIMHANS)3 Dept of Neurology, National Institute of Mental Health and Neuro Science(NIMHANS), Bangalore, India

Received May 23, 2013; Accepted November 09, 2013

ABSTRACT – Measuring neuro-haemodynamic correlates in the brain ofepilepsy patients using EEG-fMRI has opened new avenues in clinical neuro-science, as these are two complementary methods for understanding brainfunction. In this study, we investigated three patients with drug-resistantreflex epilepsy using EEG-fMRI. Different types of reflex epilepsy such aseating, startle myoclonus, and hot water epilepsy were included in thestudy. The analysis of EEG-fMRI data was based on the visual identifica-tion of interictal epileptiform discharges on scalp EEG. The convolutionof onset time and duration of these epilepsy spikes was estimated, andusing these condition-specific effects in a general linear model approach,we evaluated activation of fMRI. Patients with startle myoclonus epilepsyexperienced epilepsy in response to sudden sound or touch, in asso-ciation with increased delta and theta activity with a spike-and-slow-wavepattern of interictal epileptiform discharges on EEG and fronto-parietalnetwork activation pattern on SPECT and EEG-fMRI. Eating epilepsy wastriggered by sight or smell of food and fronto-temporal discharges werenoted on video-EEG (VEEG). Similarly, fronto-temporo-parietal involvementwas noted on SPECT and EEG-fMRI. Hot water epilepsy was triggered bycontact with hot water either in the bath or by hand immersion, and VEEGshowed fronto-parietal involvement. SPECT and EEG fMRI revealed a simi-lar fronto-parietal-octhat continuous EEG rrelated to interictal epthe neuro-haemodyn

Key words: reflex ep

19

cipital involvement. From these results, we concludeecording can improve the modelling of BOLD changesileptic activity and this can thus be used to understandamic substrates involved in reflex epilepsy.

ilepsy, EEG, fMRI, eating, hot water, startle myoclonus

mailto:[email protected]


2

RoiiotstitupwhrTniplsG2IpIsTtwfcnErfstf

M

WAudciaielrt

SwnpMissaiCavp

P

CstNeTlndpiaTblietpaifriChaopao

eflex epilepsy has always been an important “modelf study” for understanding the mechanisms involved

n triggering seizures (Chifari et al., 2004). Advancesn neuroimaging and EEG recording techniques havepened new horizons for the study of mechanisms

hat trigger seizures. Functional imaging provides highpatial resolution, but lacks the temporal resolutiono model neuronal processing. EEG provides neuronalnformation with high temporal resolution, but lackshe spatial resolution of MRI. Multimodal imagingtilizes the information provided by these two com-lementary methods, and EEG monitoring togetherith functional MRI (EEG-fMRI) is used to map neuro-aemodynamic changes with good spatio-temporalesolution (Mazziotta et al., 2001; Rosa et al., 2010).his technique, when extended to epilepsy, offers aew opportunity to localise the generators of inter-

ctal epileptiform discharges (IEDs) and capture theerfusion changes which occur using blood oxygen

evel-dependent (BOLD) contrast at time points of theeizure activity provided by EEG (Müri et al., 1998;oldman et al., 2000; Diehl et al., 2003; Mulert et al.,

005).n this study, we used simultaneous EEG-fMRI inatients with reflex epilepsy to localise generators of

EDs. Brain MRI in these cases did not reveal obvioustructural lesions which could account for seizures.he onset and duration of each IED was noted fromhe entire EEG recorded signal and this informationas fed to a general linear model (GLM) of resting

MRI in order to determine the neuro-haemodynamicorrelates of these IEDs. Since EEG-fMRI is a relativelyew technique, we validated the findings with video-EG (VEEG) and interictal and ictal SPECT findings. Theegional cerebral blood flow (rCBF) values obtainedrom interictal SPECT localise seizure focus with a sen-itivity of 75% (Menzel et al., 1995) and we correlatedhese findings with interictal BOLD changes obtainedrom EEG-fMRI.

aterials and methods

e studied three cases of reflex epilepsy.ll patients underwent EEG and brain fMRI afterndergoing VEEG, brain MRI, and SPECT imaging twoays prior to the study. A dedicated epilepsy proto-ol, which included T1MPRAGE and high-resolution

0

maging of hippocampus, was performed for brain MRIcquisition. VEEG and SPECT were recorded duringctal and interictal periods. The SPECT perfusion agent,thylcysteinate dimer (ECD), was used to detect vascu-

ar changes associated with epilepsy. The SPECT dataecorded was further processed using Lassens correc-ion to obtain the rCBF maps (Matsuda et al., 1993).

ifrl(io

ensory stimulus, specific to the type of reflex epilepsy,as given to all patients before being taken to the scan-er for EEG-fMRI investigation, during the interictaleriod. An EEG was also obtained prior to entering theRI scanner, before the actual recording of EEG-fMRI,

n order to increase the yield of obtaining IEDs duringimultaneous acquisition. EEG was visually analysed forpikes by a neurologist. Institutional ethics clearancend informed consent from the patient was obtainedn all the cases.linically, all the patients had refractory reflex epilepsynd were on medications, and seizures could be pro-oked by sensory stimulus. None of these cases werelanned for surgery.

atient details

ase 1 was a 17-year-old, right-handed boy withtartle myoclonus. He developed seizures in responseo sudden sensory stimuli. Age at onset was 2 years.eurological examination revealed blindness. Thepilepsy was triggered by sudden sound and touch.he seizure was not associated with preictal aura or

oss of consciousness or postictal tiredness. There waso family history of epilepsy. EEG showed increasedelta and theta activity with a spike-and-slow-waveattern of IEDs. Brain MRI revealed changes of hypoxic

schaemic injury in bilateral occipital lobes. Interictalnd ictal SPECT was performed 48 hours beforehand.he interictal scan revealed perfusion changes in theilateral frontal (more on the left than right side) and

eft parietal lobes (figure 1A). This was more extensiven the ictal period (figure 1B) and involved the bilat-ral frontal and parietal temporal lobes as well as thehalamus. Hypoperfusion was noted in bilateral occi-ital lobes in both ictal and interictal recordings, inccordance with MRI findings of hypoxic ischaemicnjury. The patient was taken for simultaneous EEG-MRI investigation after interictal discharges wereecorded on EEG, before entering the scanner follow-ng a sudden auditory stimulus.ase 2 was an 11-year-old, right-handed male with aistory of eating epilepsy. He had epilepsy since thege of eight years. The epilepsy was triggered by sightr smell of food. The seizure was not associated withreictal or postictal symptoms. His elder sibling hadhistory of complex partial seizures since the age

f 2 years. Video-EEG suggested left fronto-temporalnterictal discharges. Brain MRI did not reveal any


ocal lesion. SPECT images during the interictal periodevealed perfusion changes involving left frontal,eft parietal, left perirolandic, and occipital areasfigure 2A). These changes were more extensive duringctal SPECT and involved bilateral frontal, temporalccipital, and parietal lobes (figure 2B). The patient


Epileptic Disord, Vol. 16, No. 1, March 2014 21

EEG fMRI in reflex epilepsy

A B

91

17

R L

CFp1

200 µVFp2

F3

F4

C3C4

P3P4

O1

O2

F7

F8

T7

T8

P7

P8

Fz

Cz

Pz

OzFC1

FC2CP1

CP2

FC5

FC6

CP5

CP6TPg

TP10

POz

R128 R128 R128 R128

Figure 1. A case of startle myoclonus (case 1).


2

M. Sandhya, et al.

40

30

20

10

0

15

10

5

0

RLD

Figure 1. (case 1, continued) (A) Interictal SPECT showing perfusion changes (rCBF) in the bilateral frontal region, more prominent onthe left than the right side (yellow arrows; coronal image), and left parietal region (yellow arrows; axial plane and sagittal plane).(B) Ictal SPECT showing perfusion changes (rCBF) in the bilateral frontal (coronal image), temporo-parietal (sagittal image), and bilateralfronto parietal regions with thalamic involvement (axial image). (C) Interictal EEG recording during simultaneous fMRI acquisitionafter cardio-ballastic and gradient artefact corrections. Increased delta and theta activity with spike-and-slow-wave pattern of IEDs isn RI (Mp invorT spon

wieCayawtoBiiriTios

D

EEt

Tamtrwawfiw

fRSppmrt

oted. (D) Location of peaks of regional activation using EEG-fMerirolandic, left post central gyrus, and bilateral parietal lobesight thalamus.he common network revealed by EEG-fMRI and SPECT (A) corre

as taken for simultaneous EEG-fMRI recording afternterictal discharges were recorded on EEG, beforentering the scanner following an eating stimulus.ase 3 was an 18-year-old, right-handed female withhistory of hot water epilepsy since the age of 9

ears. Generalised tonic-clonic epilepsy manifestedfter bathing in hot water or by touching things thathere hot. The seizure was not associated with preic-

al or postictal symptoms. There was no family historyf epilepsy. VEEG recorded left fronto-parietal IEDs.rain MRI was normal. SPECT images during the

nterictal period revealed perfusion changes involv-ng bilateral frontal, parietal temporal, and occipitalegions (figure A). The changes were more extensiven the same areas during the ictal recording (figure 3B).he patient was taken for simultaneous EEG-fMRI

nvestigation after interictal discharges were recordedn EEG before entering the scanner, following immer-ion of a hand in hot water.

2

ata acquisition

EGEG data were recorded using a 32-channel MR compa-ible EEG system (Brain Products, Gilching, Germany).

(vpa6rv

NI stereo tactic coordinates, p<0.001 FDR uncorrected): the leftlving the superior and inferior parietal lobule, precuneus, and

ded to the left frontal parietal network.

he EEG cap consisted of 31 scalp electrodes placedccording to the international 10-20 electrode place-ent system and one additional electrode dedicated

o the electrocardiogram (ECG). Data were recordedelative to an FCz reference and a ground electrodeas located at Iz (10-5 electrode system) (Oostenveld

nd Praamstra, 2001). Data were sampled at 5,000 Hz,ith a bandpass of 0.016-250 Hz along with 50-Hz notchltering. The impedance between electrode and scalpas kept below 5 k�.

MRIesting fMRI was acquired using a 3T scanner (Skyra,iemens, Erlangen, Germany). The subject’s head wasositioned within a 20-channel head coil with foamadding to provide comfort and to minimise headovements. T1-weighted, three dimensional high

esolution imaging (MPRAGE) was performed to facili-ate localisation of fMRI activation. Echo-planar imagesEPI) using BOLD contrast were acquired and 185olumes were obtained by applying the following EPI


arameters: 34 slices, 6-mm slice thickness withoutny interslice gap, FOV of 192×192 mm, matrix of4×64, repetition time of 3,000 ms, echo time of 35 ms,efocusing pulse of 90◦, matrix of 256×256×114, andoxel size of 3×3×4 mm.




P:P

P

31s

T

S

92

4 0

R L

C

A B

Fp1100 µV

Fp2

F3

F4

C3

C4

P3

P4

O1

O2

F7

F8

T7

T8

P7

P8

Fz

Cz

Pz

Oz

FC1

FC2

CP1

CP2

FC5

FC6

CP5

CP6

TPg

TP10

POz

R128 R128 R128

Figure 2. A case of eating epilepsy (case 2).


2

M. Sandhya, et al.

10

0

5

10

0

5

D

Figure 2. (case 2, continued) (A) Interictal SPECT showing perfusion changes (rCBF) in the left frontal, left perirolandic (yellow arrow;coronal plane), left parietal (yellow arrow; axial plane), and occipital areas (yellow arrow -sagittal plane). (B) Ictal SPECT showingp ipital( n aftes al acF eral pp d mT spon

D

ERVGubfbDgffi

EFsaet(

fOdptmrt

aPtcftowaltgct

R

Ggools

C

erfusion changes (rCBF) in the bitemporal (coronal plane), occC) Interictal EEG recording during simultaneous fMRI acquisitioharp wave discharges are noted. (D) Location of peaks of regionDR uncorrected.): the left fronto-temporal (coronal plane), bilataracentral lobule, cingulate, medial frontal gyrus, and lingual anhe common network revealed by EEG-fMRI and SPECT (A) corre

ata analysis

EG data artefact removal and preprocessingaw EEG data were processed offline using Brainision Analyzer version 2 (Brain Products, Gilching,ermany). Gradient artefact correction was performedsing modified versions of the algorithms proposedy Allen et al. (2000), whereby a gradient arte-

act template was subtracted from the EEG using aaseline-corrected sliding average of 20 MR volumes.ata were then down-sampled to 250 Hz. Followingradient artefact correction, the data were correctedor cardio-ballistic artefacts. For this, the average arte-act subtraction method of Allen et al. (1998) wasmplemented using the Brain Vision Analyzer 2.

EG-informed fMRI analysisor the analysis of the resting functional data at theingle subject level, we treated the IEDs of the EEGs single trial “event-related potentials (ERPs)” in anvent-related design, considering the onset and dura-ion as explanatory variables in the GLM analysisWarbrick et al., 2009).

MRI analysis

4

ur study aimed to detect haemodynamic responseuring epileptogenic discharge in reflex epilepsyatients. The fMRI analysis was performed using sta-

istical parametric mapping (SPM8; Welcome Depart-ent of Cognitive Neurology, London). The data were

ealigned to the mean image, were then normalised tohe Montreal Neurological Institute (MNI) template,

Eoinaid

(sagittal plane), and bilateral fronto-parietal (axial plane) areas.r cardio-ballastic and gradient artefact corrections. Generalised

tivation using EEG-fMRI (MNI stereo tactic coordinates, p<0.001arietal areas (coronal plane), and mesial structures, such as the

edial occipital gyrus (axial and sagittal plane).ded to the left frontal parietal occipital network.

nd were smoothed with a Gaussian kernel of 6 mm.reprocessed data were modelled using the onsetime and the duration of each IEDs as an regressor,onvolved with a canonical haemodynamic responseunction (Friston et al., 2007). An F-contrast was used toest for BOLD signal changes related to time of onsetf the IEDs. Statistics were set at p<0.001, uncorrectedith a cluster size of 5 voxels since all the cases were

ssessed individually. The activated brain areas wereabelled after transforming the coordinates of MNIemplate to Talairach coordinates. The areas showingreatest cluster size and mean intensity (T score) wereonsidered significant for the purpose of lateralisinghe epilepsy network in a given case.

esults

ood quality EEG was obtained following pulse andradient artefact subtraction, allowing identificationf IEDs. Recordings of interictal discharges werebserved and these were considered for further ana-

ysis. Head motion artefacts were absent since all thetudies were performed during the interictal period.

ase 1: startle myoclonus


EG changes included a spike-and-slow-wave patternf IEDs, the onset and duration of which was convolved

nto the fMRI (figure 1C). Significant activations wereoted in the left perirolandic, left post central gyrus,nd bilateral parietal lobes involving the superior andnferior parietal lobule and precuneus on fMRI, withesign based on the EEG data during simultaneous




A

C

B

82

9

R L

Fp150 µV

Fp2

F3

F4

C3

C4

P3

P4

O1

O2

F7

F8

T7

T8

P7

P8

Fz

Cz

Pz

Oz

FC1

FC2

CP1

CP2

FC5

FC6

CP5

CP6

TPg

TP10

POz

R128 R128 R128

Figure 3. A case of hot water epilepsy (case 3).


2

M. Sandhya, et al.

L R

100

0

50

D

Figure 3. (case 3, continued) (A) Interictal SPECT showing perfu-sion changes (rCBF) in the bilateral frontal (yellow arrow; axialplane), biparietal (axial plane), temporal (yellow arrow; coro-nal plane), and occipital (sagittal plane) areas. (B) Ictal SPECTshowing perfusion changes (rCBF) in the bilateral frontal, pari-etal, occipital (axial and sagittal plane), and bitemporal (coronalplane) regions. (C) Interictal EEG recording during simultaneousfMRI acquisition after cardio-ballastic and gradient artefact cor-rections. Spike-wave complex IEDs are noted. (D) Location ofpeaks of regional activation using EEG-fMRI (MNI stereo tacticcilmTc

ETct

CE(ctsrmgmti

CEoc

nsgppft(

D

SUsotateiwwmfRaoeoni(apifArmpfdptdEasand subcortical neural systems, putatively involved in

oordinates, p<0.001 FDR uncorrected.): the left superior andnferior frontal gyrus, mesial structures (such as the paracentralobule), cingulate gyrus, medial frontal gyrus, and lingual and

edial occipital gyrus.he common network revealed by both EEG fMRI and SPECT (A)orresponded to the fronto-parietal occipital areas.

EG-fMRI recording in the interictal period (figure 1D).he other neuronal substrate, for which significantlusters were obtained, corresponded to the righthalamus (figure 1D).

ase 2: eating epilepsyEG showed generalised sharp wave dischargesfigure 2C). The onset and duration of each IED wasonvolved into the fMRI. Activation was noted inhe left lateral fronto-temporal lobes involving theuperior frontal gyrus, inferior frontal gyrus, supe-ior temporal gyrus, bilateral parietal region, and the

esial structures, such as the paracentral lobule, cin-ulate gyrus, caudate, medial frontal gyrus, lingual, andedial occipital gyrus (figure 2D), with design based on

he EEG data during simultaneous EEG-fMRI recording

6

n the interictal period.

ase 3: hot water epilepsyEG changes noted included the spike-wave complexf the IEDs, the onset and duration of which wasonvolved into the fMRI (figure 3C). Activations were

reEtur

oted in the left fronto-parietal structures involvinguperior frontal gyrus and right superior temporalyrus. Medial structures involved were the parahip-ocampal gyrus, superior parietal lobule, inferiorarietal lobule, posterior cingulate, and precuneus on

MRI, with design based on the EEG data during simul-aneous EEG-fMRI recording in the interictal periodfigure 3D).

iscussion

ystem epilepsies are a recently described entity.nlike a classic seizure which spreads from a circum-

cribed region and propagates sequentially along oner more neural pathways to other brain areas, sys-

em epilepsy describes the susceptibility of a systems a whole, although it may be possible to iden-ify some trigger areas within the system (Avanzinit al., 2012). There may be symmetric or assymetric

nvolvement of cortical and subcortical structures withidespread multifocal or focal involvement. The net-ork involved depends on the age and degree ofaturity of the brain networks and can manifest as

ocal or generalised epilepsy (Capovilla et al., 2009).eflex epilepsy is one such entity in system epilepsynd can provide unique insight into the ictogenesisf human epilepsy and networks involved (Avanzinit al., 2012). Reflex seizures can be either generalisedr focal, with or without impairment of conscious-ess, and may be genetically inherited or acquired,

diopathic, cryptogenic or supported by a brain lesionStriano et al., 1989). Three types of reflex epilepsyre described; those induced by simple stimuli, com-lex stimuli, or higher cerebral functions. Functional

maging techniques such as PET, SPECT, and EEG-MRI provide the potential to image epileptic activity.bnormal EEG observed in patients with epilepsy is

eflected by interictal epileptiform discharge, com-only called the “epileptic spike”. The SPECT imaging

rotocol includes both interictal and ictal imaging. EEGMRI image acquisition is possible only during IED,ue to gross patient motion, inherent during the ictalhase. Both EEG-fMRI and SPECT can be used to reflect

he haemodynamic correlates of ictal and interictalischarges non-invasively.EG-fMRI exhibits good spatio-temporal resolutionnd can therefore be profitably used to study thetructure and function of the distributed cortical


eflex epilepsy, even in cryptogenic cases (Boucousist al., 2012).CD is a SPECT perfusion agent which is lipophilic andherefore diffuses into the normal brain and can besed to study perfusion changes during epilepsy. TheCBF values obtained after post-processing of SPECT



data during ictal and interictal phase reflect the bloodflow changes associated with epilepsy.In this study, the aim was to validate EEG-fMRI asa potential clinical tool for functional imaging withhigh spatio-temporal resolution to detect both corticaland subcortical networks involved in reflex epilepsy.The findings were correlated with interictal and ictalSPECT perfusion imaging and ictal VEEG for validation.Three types of reflex epilepsy were considered; startlemyoclonus (somatosensory evoked epilepsy; simplestimuli) and eating and hot water epilepsy (complexstimuli evoked epilepsy). The foci were cryptogenic onMRI.In startle epilepsy, seizures are mostly axial tonicand provoked by a sudden and unexpected stimu-lus. Structural imaging has shown perinatal hypoxicinjury as the most common lesion associated withstartle epilepsy (Alajouanine and Gastaut, 1955). Manytheoretical and practical factors, including diagnostic,genetic, and pathophysiological issues still remainunresolved for both these entities. These issues canonly be addressed by using robust imaging techniquesin order to understand the pathophysiology, with goodspatio-temporal resolution. Functional neuroimagingstudies in startle epilepsy have implicated an inter-action with the mesial fronto-parietal network, involv-ing the supplementary motor area, perirolandic cortex,and precuneus (Fernández et al., 2011). In our study,

Table 1. Summary of the findings in reflex epilepsy usinbased on our study an

Type of reflexepilepsy

VEEG SPECT

Startlemyoclonus

Not available Bilateral frontal (leftmore than right), leftparietal lobe, andbilateral thalamii(ictal scan)

Eating Leftfronto-parietal

Left frontal-parietalperirolandic area andoccipital lobes

Hot water Leftfronto-temporal

Bilateralfronto-parietaltemporal occipitallobes

aSEspDi(diooaTp(chfsabf(Hshr


similar network was involved using EEG-fMRI andPECT during the interictal period (table 1).pilepsy provoked by eating is a rare entity and the ictalemiology differs from patient to patient. The complexartial type of seizure is most commonly described.iffuse cerebral damage is often noted on MRI and

t may coexist with generalised or absence seizuresLoreto et al., 2000). Reproduction of these seizuresuring EEG recording is often difficult as the stimulus

s frequently complex, involving different componentsf eating, such as the sight of food, proprioceptive,lfactive or gustative stimulations, chewing, salivation,nd gastric distension of eating (Cirignotta et al., 1977).wo foci have been described in the amygdala or theerirolandic location, as demonstrated by EEG studies

Remillard et al., 1998). A high incidence of familiallustering is known and bulky meals rich in carbo-ydrates have been postulated as possible triggering

actors (Nagaraja and Chand, 1984). The case pre-ented here was eating epilepsy triggered by olfactivend gustative stimulation, and the network implicated,

27

g EEG-fMRI and correlation with VEEG and SPECTd other studies.

EEG-fMRI Other studies

Left fronto-parietal lobes(post central gyrus,perirolandic, precuneus,superior, and inferiorparietal lobule) and rightthalamus.

Mesial fronto-parietalnetwork (supplementarymotor area, perirolandiccortex, and precuneususing fMRI) (Boucousiset al., 2012).

Left frontal temporal andbilateral parietal lobes, andmesial structures such asparacentral lobule,cingulate, caudate, medialfrontal gyrus, and lingualand medial occipital gyri

Amygdala or in theperirolandic regionusing EEG (Loreto et al.,2000).

Left fronto-parietaltemporal occipital lobesincluding mesial structures(parahippocampal gyrus,superior parietal lobule,posterior cingulate, andprecuneus)

Left temporo-occipitalarea using EEG (Nagarajaand Chand, 1984).

ased on EEG-fMRI, VEEG, and SPECT, was the leftrontal parietal network and temporo-occipital areastable 1).ot water epilepsy is most frequently reported in

outhern India due to the peculiar habit of pouringot water over the head. This form of epilepsy is ageelated, more frequent in males, usually self-limited,


2

M

aPrwathcewtniTtpIatotafc

C

Tttiasotsa

AOmSif(IT

R

AsI1

Afip1

AiM

At2

Boa

Ccg2

CWm

Cc

Dsmf

Fit

Fti

Gi2

La2

Lat9

MNuE

M

. Sandhya, et al.

nd frequently familial (Satishchandra et al., 1988).atients are technically difficult to study in the labo-atory. Lisovoski et al. (1992) reported a case of hotater epilepsy in which ictal EEG, recorded duringhot bath, showed focal epileptic discharges in the left

emporo-occipital area and interictal SPECT showedypometabolism in the same cerebral region. Ourase is unique in that firstly, the patient develop-d significant IED even on immersion of the hand in hotater and this was replicated during the recording in

he laboratory, and secondly, the left temporo-occipitaletwork was involved, based on all three modalities,

.e. VEEG, EEG-fMRI, and SPECT (table 1).his study is one of the first to describe the simul-aneous role of EEG and fMRI in the study ofathophysiology of different types of reflex epilepsy.

t is particularly useful in reflex epilepsy because, inddition to being a rare entity, in most of the caseshe lesion is cryptogenic on MRI. Our study has notnly demonstrated the network involved in each of

hese rare epilepsies in individual cases, but has alsottempted to lateralise the onset zone by using per-usion imaging such as SPECT and VEEG findings fororrelation.

onclusion

he data discussed in this study supports the hypo-hesis that EEG-fMRI is a useful clinical tool to studyhe patho-physiological mechanisms and networksnvolved in epilepsy, particularly system epilepsy, suchs reflex epilepsy, of which the pathophysiology iso far not understood. These developments can helppen up new therapeutic possibilities. Further studies

o validate and assess possible gains in sensitivity andpecificity of the method are required and should beddressed in larger groups of patients. �

cknowledgements and disclosures.ur Cognitive Neuroscience Center is supported by the Depart-ent of Science and Technology (DST), India. We thank Dr

hobini L Rao (Dept. Clinical Psychology, NIMHANS, India) fornspiring us to carry out the research of simultaneous EEG-MRI and multimodal neuroimaging. We thank our radiographersDept. Neuroimaging and Interventional Radiology, NIMHANS,ndia) for their support during data collection.he authors have no conflicts of interests to disclose.

8

eferences

lajouanine T, Gastaut H. Synkinesis-startle and epilepsytartle triggered by unexpected sensory and sensitive factors.. Anatomical and clinical data on 15 cases. Rev Neurol (Paris)955; 93: 29-41.

rtB

Min1

llen PJ, Polizzi G, Krakow K, Fish DR, Lemieux L. Identi-cation of EEG events in the MR scanner: the problem ofulse artifact and a method for its subtraction. Neuroimage998; 8: 229-39.

llen PJ, Josephs O, Turner R. A method for removing imag-ng artifact from continuous EEG recorded during functional

RI. Neuroimage 2000; 12: 230-9.

vanzini G, Manganotti P, Meletti S, et al. The sys-em epilepsies: a pathophysiological hypothesis. Epilepsia012; 53: 771-8.

oucousis SM, Beers CA, Cunningham CJB, et al. Feasibilityf an intracranial EEG-fMRI protocol at 3T: risk assessmentnd image quality. Neuroimage 2012; 63: 1237-48.

apovilla G, Berg AT, Cross JH, et al. Workshop Report: con-eptual dichotomies in classifying epilepsies: partial versuseneralized and idiopathic versus symptomatic (April 18-20,008, Monreale, Italy). Epilepsia 2009: 1645-9.

hifari R, Piazzini A, Turner K, Canger R, Canevini MP,olf P. Reflex writing seizures in two siblings with juvenileyoclonic epilepsy. Acta Neurol Scand 2004; 109: 232-5.

irignotta F, Marcacci G, Lugaresi E. Epileptic seizures pre-ipitated by eating. Epilepsia 1977; 18: 445-9.

iehl B, Salek-haddadi A, Fish DR, Lemieux L. Mapping ofpikes, slow waves, and motor tasks in a patient with malfor-ation of cortical development using simultaneous EEG and

MRI. Magn Reson Image 2003; 21: 1167-73.

ernández S, Donaire A, Maestro I, et al. Functional neuro-maging in startle epilepsy: involvement of a mesial fron-oparietal network. Epilepsia 2011; 52: 1725-32.

riston KJ, Ashburner JT, Kiebel SJ, Nichols TE, Penny WD. Sta-istical parametric mapping: the analysis of functional brainmages. London: Academic Press, 2007.

oldman RI, Stern JM, Engel J, Cohen MS. Acquir-ng simultaneous EEG and functional MRI. Neurophysiol000; 111: 1974-80.

isovoski F, Prier S, Koskas P, et al. Hot-water epilepsy inn adult: ictal EEG, MRI and SPECT features. Seizure 1992; 1:03-6.

oreto V, Nocerino C, Striano P, D’Aulos F, Boccella P, Stri-no S. Eating epilepsy. Heterogeneity of ictal semiology:he role of video-EEG monitoring. Epileptic Disord 2000; 2:3-8.

atsuda H, Tsuji S, Shuke N, Sumiya H, Tonami N, Hisada K.oninvasive measurements of regional cerebral blood flowsing technetium-99m hexamethylprophylene amine oxime.ur J Nucl Med 1993; 20: 391-401.

azziotta J, Toga A, Evans A, et al. A probabilistic atlas andeference system for the human brain: International Consor-


ium for Brain Mapping (ICBM). Philos Trans R Soc Lond Biol Sci 2001; 356: 1293-322.

enzel C, Hufnagel A, Grünwald F, et al. The relevance ofnterictal rCBF brain SPECT in temporal lobe epilepsy: diag-ostical value and effects of spatial resolution. Ann Nucl Med995; 9: 215-23.


E

Mom4

MReM

N1

OtN

Re

Rmu

SSe

Src

ulert C, Jäger L, Propp S, et al. Sound level dependencef the primary auditory cortex: simultaneous measure-ent with 61-channel EEG and fMRI. Neuroimage 2005; 28:

9-58.

üri RM, Felblinger J, Rösier KM, Jung B, Hess CW, Boesch C.ecording of electrical brain activity in a magnetic resonancenvironment: distorting effects of the static magnetic field.agn Reson Med 1998; 39: 18-22.

agaraja D, Chand RP. Eating epilepsy. Clin Neurol Neurosurg

pileptic Disord, Vol. 16, No. 1, March 2014

984; 86: 95-9.

ostenveld R, Praamstra P. The five percent electrode sys-em for high-resolution EEG and ERP measurements. Clineurophysiol 2001; 112: 713-9.

emillard GM, Zifkin BG, Andermann F. Seizures induced byating. Adv Neurol 1998; 75: 227-40.

IG4

Wayt


osa MJ, Kilner J, Blankenburg F, Josephs O, Penny W. Esti-ating the transfer function from neuronal activity to BOLD

sing simultaneous EEG-fMRI. Neuroimage 2010; 49: 1496-509.

atishchandra P, Shivaramakrishana A, Kaliaperumal VG,choenberg BS. Hot-water epilepsy: a variant of reflexpilepsy in southern India. Epilepsia 1988; 29: 52-6.

triano S, Meo R, Bilo L, Perrone M. An attempt to classifyeflex seizures in accordance to recent proposals for classifi-ation of epilepsies. In: Reflex seizures and reflex epilepsies.

29

nternational colloquium on epileptic and reflex seizures.eneva (Switzerland): éditions Medecine and Hygiene, 1989:

89-93.

arbrick T, Mobascher A, Brinkmeyer J, et al. Single-trial P3mplitude and latency informed event-related fMRI modelsield different BOLD response patterns to a target detectionask. Neuroimage 2009; 47: 1532-44.

Understanding the pathophysiology of reflex epilepsy using simultaneous EEG-fMRI

Documents

Transcript of Understanding the pathophysiology of reflex epilepsy using simultaneous EEG-fMRI