FAST_6454B8 的标签_百度知道
FAST_6454B8 的标签
我有更好的答案
wifi路由器的名字
其他类似问题
为您推荐：
等待您来回答
下载知道APP
随时随地咨询
出门在外也不愁文档分类：
在线文档经过高度压缩，下载原文更清晰。
淘豆网网友近日为您收集整理了关于EEG-triggered functional MRI of interictal epileptiform activity in patients with partial seizures的文档，希望对您的工作和学习有所帮助。以下是文档介绍：EEG-triggered functional MRI of interictal epileptiform activity in patients with partial seizures Brain (1999), 122, EEG-triggered functional MRI of interictalepileptiform activity in patients with partialseizuresK. Krakow,1,4 F. G. Woermann,1,4 M. R. Symms,1,4 P. J. Allen,3 L. Lemieux,1,4 G. J. Barker,2J. S. Duncan1,4 and D. R. Fish1,31The Epilepsy Research Group and 2NMR Research Unit, Correspondence to: Dr Karsten Krakow, MD, MRI Unit,Department of Clinical Neurology, Institute of N(来源：淘豆网[/p-7498645.html])eurology, National Society for Epilepsy, Chalfont St Peter, Gerrards3Department of Clinical Neurophysiology, The National Cross, Bucks SL9 0RJ, UKHospital for Neurology and Neurosurgery, London and E-mail: kkrakow@ion.ucl.ac.uk4National Society for Epilepsy, Chalfont St Peter, UKSummaryEEG-triggered functional MRI (fMRI) offers the potentialto localize the generators of scalp EEG events, suchas interictal epileptiform discharges, using a biologicalm(来源：淘豆网[/p-7498645.html])easurement as opposed to relying solely on modellingtechniques. Although recent studies have demonstratedthese possibilities in a small number of patients, widerapplication has been limited by concerns about patientsafety, severe problems due to pulse-related artefactobscuring the EEG trace, and lack of reproducibilitydata. We have systematically studied and resolved theissues of patient safety and pulse artefact and now reportthe application of the(来源：淘豆网[/p-7498645.html]) technique in 24 experiments in 10consecutive patients with localization-related epilepsy andfrequent interictal epileptiform discharges (spikes or spikewave). At least two experiments were performed for eachpatient. In each experiment, 10- or 20-slice snapshotKeywords: functional MRI; EEG; interictal epi epilepsy surgeryAbbreviations: BOLD blood oxygen level- EPI fMRI functional MRI; SPE(来源：淘豆网[/p-7498645.html])CTsingle photon puted tomographyIntroductionInterictal epileptiform discharges recorded by scalp EEG arethe mainstay for classifying types of epilepsy, and theirlocalization has an important role in the presurgical evaluationof drug-resistant patients (Gilliam et al., 1997). Knowledgeof the underlying generators of these EEG events, however,is still limited. Due to their restricted spatial sampling andthe ‘inverse problem’(working back from distant (来源：淘豆网[/p-7498645.html])scalppotentials to hypothesize about the likely sites of theirgenerators), neither EEG nor oencephalography candirectly identify these generators (Ebersole, 1998). On theother hand, the low temporal resolution of PET and single Oxford University Press 1999gradient-echo planar images were acquired ~3.5 s after asingle typical epileptiform discharge (activation image)and in the absence of discharges (control image). Between21 and 50 epileptiform disch(来源：淘豆网[/p-7498645.html])arges were sampled in eachexperiment. The signicance of functional activation wastested using the t test at 95% condence on a pixel-by-pixel basis. Six of the 10 patients showed reproduciblefocal changes of the blood oxygen level-dependent (BOLD)signal, which occurred in close spatial relationship to themaximum of the epileptiform discharges in the concurrentEEG. No reproducible focal BOLD signal changes wereobserved in the remaining four patients. (来源：淘豆网[/p-7498645.html])In conclusion,EEG-triggered fMRI is now a sufciently developedtechnique to be more widely used in clinical studies,demonstrating that it can reproducibly localize the brainareas involved in the generation of spikes and spike wavein epilepsy patients with frequent interictal discharges.photon puted tomography (SPECT) prevents theinvestigation of brain activation linked to brief epileptiformdischarges.In contrast, case reports of patients with localiz(来源：淘豆网[/p-7498645.html])ation-related epilepsy have shown recently that functional MRI(fMRI) may allow the detection of local changes in bloodoxygenation associated with subclinical epileptic seizures(Jackson et al., 1994) and interictal epileptiform discharges(Warach et al., 1996; Seeck et al., 1998; Symms et al., 1998).The acquisition of MRI linked to brief subclinical eventsrequires the recording of the EEG during the MR scanning1680 K. Krakow et al.procedure. It has be(来源：淘豆网[/p-7498645.html])en shown that the EEG can be recordedinside an MR scanner with sufcient quality to detect high-amplitude discharges and to trigger fMRI acquisitions afterthese events (Ives et al., 1993; Huang-Hellinger et al., 1995).However, safety issues and EEG artefacts due to icelds in the MR scanner which may obscure the EEG tracehave limited the method so far and precluded its widerapplication (Ives et al., 1993; Huang-Hellinger et al., 1995).We have establis(来源：淘豆网[/p-7498645.html])hed a protocol to ensure patient safety duringEEG recording (Lemieux et al., 1997) and developed amethod for the on-line subtraction of pulse artefact ( al., 1999), which has been regarded as the most signicantEEG artefact inside the MR scanner (Ives et al., 1993;Huang-Hellinger et al., 1995; Allen et al., 1999).We used this new recording technique to monitor the EEGof patients with partial seizures undergoing MRI and triggeredultra-fast snapshot multislice echo planar imaging (EPI) bloodoxygen level-dependent (BOLD) fMRI acquisitions aftersingle epileptiform discharges (spike or spike wave) wereidentied in the on-line EEG. The site of the fMRI pared with the focus of previous interictal scalpEEG recordings and, if available, invasive and ictal EEGrecordings. All patients were studied at least twice on asions in order to investigate the reproducibility of thefMRI results.MethodsPatientsTen consecutive patients (seven male, three female, medianage 28.5 years, range 22–48 years) with a conrmed diagnosisof medically intractable localization-related epilepsy werestudied. The clinical data of the patients are summarized inTable 1. The study was approved by the mittee ofthe National Hospital for Neurology and Neurosurgery andall patients gave informed consent. The patients showedfrequent stereotyped focal epileptiform discharges in previousroutine 20-channel scalp EEG recordings with an average ofat least one epileptiform discharge per minute. Patients withless frequent, generalized or multifocal interictal dischargeswere excluded from the study. Five patients underwentpresurgical evaluation prior to the study, including ictal video-EEG recording (n 5) and electrocorticography (n 1).EEG recordingThe EEG was recorded in the MR scanner using the followingprocedure. Standard Ag/AgCl disk electrodes were appliedon the s these had 15 k current-limitingcarbon resistors tted adjacent to each electrode (Lemieuxet al., 1997). The electrodes were connected to a non-ferrousheadbox (developed in-house) placed at the entrance to thebore of the . The headbox was connected to aNeurolink Patient Module (Physiometrix, N.Billerica, Mass.,USA), which digitizes and transmits the EEG signal out ofthe scanner room via a bre optic cable to the NeurolinkMonitor Module, which reconstructs the analogue EEGsignals. These were then recorded using a digital EEGrecording system (sample rate 200 Hz, bandwidth 0.12–50 Hz).For each experiment, 12 electrodes were applied to scalppositions FP1/FP2, F7/F8, T3/T4, T5/T6, O1/O2, Fz and ording to the 10/20 system. In addition, two precordialECG channels were recorded to facilitate pulse artefactsubtraction (75 k current-limiting resistors were tted toeach ECG electrode) (Allen et al., 1999). The EEG datawere digitally remontaged and displayed to show bitemporalchains. In eight patients, on-line pulse artefact subtractionsoftware was used to aid visual detection of the epileptiformdischarges (Fig. 1). This method subtracts an averaged pulseartefact waveform calculated for each electrode during theprevious 10 s. Technical details have been describedelsewhere (Allen et al., 1999).fMRI acquisition and processingfMRI was performed on a 1.5 Tesla Horizon EchoSpeedMRI scanner (General Electric, Milwaukee, Wis., USA) usingsnapshot gradient-echo EPI [TE (echo time) 40 ms, 24cm eld of view]. In the rst studies of patients 1 and 2, 10contiguous 5 mm slices were acquired with a 128 128matrix. Patient 1 was studied in axial and coronal orientations.Patient 3 (with a proposed epileptogenic zone in the mesialtemporal lobe) was studied in coronal orientation to bettervisualize mesial temporal lobe structures and with a reducedmatrix size (64 64) to reduce the susceptibility artefactsdue to the presence of air in nearby sinuses. For all otherexperiments, axial acquisition with 20 contiguous 5 mmslices with a 64 64 matrix was performed (Table 2). Theacquisition time was 3.5 s for 10 slices and 4.5 s for 20slices. For all studies based on a 64 64 matrix, additionalhigh-resolution multi-shot EPI images [matrix 256 256,16 shots, repetition time 3 s, all other parameters as fMRIdata] were acquired. These images have geometric distortionssimilar to the fMRI data and were used as anatomicalreferences for the fMRI data.Images were acquired after ‘activation’ and ‘control’states, dened by visual inspection of the on-line EEG.The activation state was dened as a single stereotypedepileptiform discharge (spike or spike wave). As the peakblood oxygenation level change detected by fMRI occurs~4–7 s after the onset of brain activity (Hennig et al., 1995;Rosen et al., 1998), a delay of ~3.5 s between observation ofthe discharge and image acquisition was applied. Acquisitionsstarted 3 or 4 s after the EEG event and acquisitionsfollowing equivocal activation or control states were excludedfrom the statistical analysis ( 7% of acquisitions in allstudies). Control images were acquired after periods of atleast 10 s of background EEG activity without epileptiformactivity. Image acquisition was performed non-periodicallywith activation and control images interleaved, depending onInterictal EEG-triggered fMRI 1681Table 1 Clinical, EEG and MRI data for the patientsPatient Age, duration A seizure type Interictal EEG Ictal EEG (onset) Structural MRI(years), sex1 48, 7 F Chronic encephalitis, left SW, left mid-temporal Widespread over Mild outer left he SPS, CPS, left hemisphere atrophySGTCS2 22, 22 M C CPS, Sp, right temporal and – Nodular heterotopia of rightSGTCS parasagittal central region and medialparietal3 26, 10 F H CPS, SW, left anterior-temporal Left sphenoidal Left temporal lobectomy.SGTCS electrode Sclerosis of residual(preoperative) hippocampus4 29, 14 M Low-grade astrocytoma SW, left frontal Widespread over Focal scarring of left middle(surgical treatment 11 years left hemisphere frontal gyrusago and radiotherapy 3 yearsago); SPS, CPS5 33, 19 M Infantile febrile seizure/ Sp, ShW, left mid- No lateralization Left-sided hippocampalh CPS temporal sclerosis6 29, 18 F Dysplastic neuroepithelial Sp, right anterior, mid- No lateralization Lesion of
CPS temporal temporal gyrus7 25, 18 M H SPC, CPS SW left temporo-central – Ischaemic brain damage ofleft post. MCA area8 43, 27 M CPS Sp, ShW right mid- – Without ndingtemporal9 24, 24 M C CPS, Sp, left anterior, mid- – Bilateral malformation ofSGTCS temporal cortical development10 28, 22 M C CPS, SGTCS SW, bilateral occipital, – Without ndingleft
SGTCS secondary generalized
S SPS siShW SW spike and slow wave.the sequence of the EEG events. An interval of at least 15 swas established between essive acquisitions to ensurethe same T1 weighting for each acquisition.Because of hardware restrictions, the number of time-points was limited to 98 per study. This led to a maximumacquisition of 49 activation and 49 control states, as equalnumbers of activation and control states were used for thestatistical analysis. The typical total scanning time was 60–90 min depending on the frequency of EEG events.The images were registered (Bullmore et al., 1996) andthen the Stimulate software (Strupp, 1996) was used toperform a two-tailed t test at the 95% condence levelbetween the activation and control images on a pixel-by-pixel basis to determine signicantly activated regions. In-plane clustering of three pixels was used to remove smalland scattered activated regions that were unlikely to representgenuine brain activity. In addition, regions of activation thatwere not evident in adjacent locations on at least twocontiguous slices were rejected.ResultsIn all 24 experiments the EEG quality was sufcient to detectactivation and control periods reliably throughout the study,although in 19 experiments (eight of the patients) pulseartefact subtraction was necessary to achieve good EEGquality (Fig. 1). Epileptiform discharges recorded inside thescanner had a localization, amplitude and congurationsimilar to those in previous recordings under routineconditions. Image acquisition for the activation states wasstarted on average 3.53 s (SD 0.24) after the epileptiformdischarges were observed in the EEG. No clinical orelectroencephalographic manifestations of ictal urred during the scanning procedure in any experiment.None of the patients reported fort or other adverseevents due to the EEG recording procedure during theexperiments.In six out of 10 patients, a signicant focal activation wasseen in the fMRI data. Signicant activations were denedas being contiguous over at least two slices and reproduciblein repeated studies. Reproducibility of activations was denedas activated voxels that were located in the same cerebrallobe and at least partly overlapped between studies. Asrepeated studies were acquired and registered independentlywithout spatial normalization, co-registration of these studiescould not be performed and the interstudy anatomicalcorrelation between activated areas was assessedquantitatively by visual inspection. In all six patients, thesereproducibility criteria were met only by a single activatedarea, which showed co-localization with the EEG spike/spikewave focus in all cases (dened by concordant lateralizationof the activated cortical area and the maximum of the1682 K. Krakow et al.Fig. 1 Patient 2. EEG recording inside the MR scanner. The same EEG sequence with a low-amplitude spike over the right hemispherewithout (A) and with (B) pulse artefact subtraction. The spike is clearly detectable only with applied pulse artefact subtraction. 3.5 safter the spike, the image acquisition artefact of a 20 slice EPI acquisition occurs.Interictal EEG-triggered fMRI 1683Table 2 Results of EEG recordings inside the MR scanner and fMRI activationsPatient Study no. Orientation No. of Matrix size No. of EEG results* fMRI results No. ofslices spikes Localization slices1 1.1 Axial 10 128 128 49 SW; T3 Left superior 31.2 Axial 10 128 128 45 184.2 (SD 33.1) V temporal lobe 31.3 Coronal 10 128 128 45 51.4 Axial 20 64 64 46 52 2.1 Axial 10 128 128 47 Sp; T4, T6 Right parietal 62.2 Axial 10 128 128 42 84.2 (SD 15.4) V lobe, within 22.3 Axial 20 64 64 47 lesion 33 3.1 Coronal 10 64 64 47 SW; T3 Left mesial 23.2 Coronal 10 64 64 46 168.1 (SD 26.3) V temporal lobe 34 4.1 Axial 20 64 64 37 SW; F7, T3 Left posterior 54.2 Axial 20 64 64 21 187.8 (SD 26.3) V frontal lobe and –4.3 Axial 20 64 64 40 superior temporal 3lobe, adjacent topostsurgical damage5 5.1 Axial 20 64 64 39 Sp, ShW; T3 No activation –5.2 Axial 20 64 64 45 68.7 (SD 14.2) V –6 6.1 Axial 20 64 64 34 Sp; T4, T6 Right perisylvian region, 36.2 Axial 20 64 64 48 118.3 (SD 13.8) V adjacent to lesion 37 7.1 Axial 20 64 64 26 SW; T5, O1 No activation –7.2 Axial 20 64 64 47 92.9 (SD 6.9) V –8 8.1 Axial 20 64 64 49 S, S T4 No activation –8.2 Axial 20 64 64 49 48.2 (SD 4.4) V –9 9.1 Axial 20 64 64 49 Sp; T3 Left temporal 49.2 Axial 20 64 64 46 67.8 (SD 16.9) V (rst experiment –only)10 10.1 Axial 20 64 64 49 SW; O1, T5, O2 Left occipital lobe 410.2 Axial 20 64 64 48 127.2 (SD 15.4) V 8*EEG recording inside the scanner: morphology of ep electrode(s) showing maximum amplitude (SSW ShW sharp wave); mean amplitude of all epileptiform discharges used to trigger fMRI acquisitions.epileptiform discharge). Details of the localization and extentof the signicant activations for all studies are given in Table2. Figures 2–4 present typical examples of activation mapsof individual fMRI studies, overlaid on high-resolution EPIimages. Areas of signicant signal increase during theactivation state are overlaid in red, representing a percentagesignal increase of 1–2%. Figure 3 gives an example of thethrough-slice contiguity of activations, and Fig. 4 presentscorresponding slices of repeated studies to provide a typicalexample of the reproducibility of activations.In some experiments several small areas appeared activatedin addition to the presumed epileptic focus. These areas couldbe distinguished from the activation concordant to the EEGfocus by (i) showing no through-slice contiguity and (ii) notbeing reproducible between studies. These signal changeswere typically localized at CSF/brain tissue boundaries,mainly at the brainstem, the interhemispheric ssure, thelateral ventricles and the frontal and occipital poles ofthe brain.The scalp EEG electrodes caused small signal dropouts inthe echo planar images, affecting mainly the scalp and skullsignal. Occasionally, the artefact also intruded into parts ofthe cortex, but not to an extent that promise theareas showing fMRI activation.In three patients (patients 5, 7 and 8), no signicantactivation was found in either of the repeated experiments.In these patients the amplitude of the interictal epileptiformactivity was lower [mean amplitude 69.9 V (SD 18.3)]compared with the patients who were studied essfully[mean amplitude 144.9 V (SD 37.9)]. In addition, in twoof these patients (patients 5 and 8) the morphology of theepileptiform activity was more variable than in the otherpatients, and not only spikes but also sharp waves wereused to trigger fMRI acquisitions. Patient 7, however, hadstereotyped spike-wave discharges with a mean amplitude of92.9 V (SD 6.9) throughout the experiment, which werenevertheless not associated with an fMRI activation. Patient9 showed a clear fMRI activation in the rst experiment,which was not reproducible in the repeated experiment.Patient 4 was studied three times, and showed a signicantfMRI activation in the rst and third experiments only, inwhich 37 and 40 epileptiform discharges, respectively, were1684 K. Krakow et al.Fig. 2 Activation map of an individual fMRI experiment overlaidon a high-resolution echo planar image of patient 2 (thecorresponding EEG is displayed in Fig. 1). The activation isadjacent to the cleft in the lateral part of the nodular heterotopia.This result was shown in all three experiments.included in the statistical analysis as activation states. Incontrast, the second experiment, comprising only 21epileptiform discharges, did not show any activation. Inour study, the smallest number of activations leading to asignicant fMRI activation was 34. In most of the experimentsshowing a positive result, 45–49 activations were sampled(Table 2).DiscussionWe found EEG-triggered EPI BOLD fMRI to be a practicableand robust method in the evaluation of epilepsy patients withfocal epileptiform discharges. In all 24 experiments, we wereable to obtain a good-quality EEG, detect spontaneousinterictal epileptiform discharges on-line and to trigger EPIBOLD acquisitions after these events.The MR image quality was not promisedby the EEG recording and in seven out of 10 patients wefound focal MR signal increases associated with the focalepileptiform discharges seen in the concurrent EEG. Theseactivations showed co-localization with the focus seen inroutine scalp EEG and in the EEG recording during theexperiment, and were reproducible in six cases. Additionalelectrocorticography was performed in one patient (patient1), and it conrmed the co-localization between interictalepileptiform activity and fMRI activation. In the three patientswith fMRI activation and previous lateralizing ictal EEGrecordings (patients 1, 3 and 4; Table 1), the ndings wereconcordant. Activation was seen in patients with differentunderlying pathologies (chronic encephalitis, hippocampalsclerosis, cortical dysgenesis, tumour). In patients with focallesions, the activation overlapped or was adjacent to thelesion (patients 2, 4 and 6; Fig. 2). Patient 3 had undergoneprevious epilepsy surgery with an anterior lobectomy of theleft temporal lobe without improving the frequency ofseizures. In this case, the fMRI revealed an activation inthe remaining mesial temporal lobe, in keeping with anepileptogenic zone beyond the previous resection (Fig. 3).The activation of deep temporal structures is remarkable asit is correlated with epileptiform discharges recorded by scalpEEG. This requires propagation of the epileptiform activityto a larger supercial cortical area. An activation solely indeep structures might suggest that fMRI more readilyidentied the site of primary spike generation. This hypothesiswould be in keeping with the result for patient 10 (Fig. 4). Inthis case the scalp EEG showed bilateral occipital discharges,while the activation map revealed unilateral occipitalactivation on the side of EEG predominance. The possibilitythat the site of the primary generator of epileptic activityis associated with different metabolic and pared with brain areas involved in thepropagation of this activity requires further study, given itspotential clinical relevance.Investigations of epileptic foci in humans have been limitedhitherto by the low spatial or temporal resolution of thediagnostic tools available. Due to their restricted spatialsampling and the insoluble inverse problem, neither EEGnor oencephalography can directly localize the sourceof epileptic activity. PET and SPECT studies have shownincreased blood ow and metabolism in the region of theseizure focus during ictal events (Engel et al., 1983; Leeet al., 1986) and, in contrast, decreased blood ow andmetabolism during the interictal state (Engel et al., 1982).Because of their low temporal resolution, however, thesemethods sample activity continuously over a prolonged periodof time, and hence cannot investigate the changes in bloodow and oxygenation related to single epileptiformdischarges. By time-locking the fMRI acquisition to singleEEG events, our study conrms the results of previous casereports (Warach et al., 1996; Seeck et al., 1998; Symmset al., 1998) that EEG-triggered fMRI can identify brainactivation associated with subclinical discharges with a highspatial resolution pletely non-invasively.Methodological considerationsMethodological limitations of spike-triggered fMRI arecaused by genuine BOLD imaging characteristics. Firstly,after a brief neuronal activation, the BOLD contrast signalchanges start to increase ~2 s after the stimulus, peak afterInterictal EEG-triggered fMRI 1685Fig.3Patient3.Activationmapofanindividualstudy.Thetwocontiguousslicesshowingactivationoftheleftmesialtemporallobearedisplayed.1686 K. Krakow et al.Fig. 4 Patient 10. Corresponding slices of the activation maps of the rst (left) and second (right) study are displayed to give a typicalexample of the reproducibility of the results. The same area in the left occipital lobe was activated in both studies.4–7 s and last ~10 s with high variability (Hennig et al.,1995; Rosen et al., 1998). These dynamics prevent thedistinction between regions that are sequentially activatedwithin a few seconds or even fractions of a second. Hence,in widespread (e.g. patient 4) or multifocal activation theprimary source of the discharge cannot be determined.Secondly, the low signal-to-noise ratio of BOLD imagingrequires sampling of activation and control states. Using a1.5 Tesla scanner, we found that at least 30 epileptiformdischarges had to be sampled to obtain an activation clearlydistinguishable from noise, even when low-resolution images(64 64 matrix) with a relatively high signal-to-noise ratio(~100) were used. This limits the applicability of this methodto patients with a high frequency of interictal epileptiformdischarges, and the duration of the study is highly dependenton the frequency of appropriate EEG events. There is atrade-off whereby increasing the number of events sampledimproves signal-to-noise ratio but requires a prohibitivescanning time and may increase misregistration problemscaused by patient movement.As epileptiform discharges are generated unpredictably,the MRI data were acquired in a non-periodic manner,with interleaved sampling of activation and control pared with continuous image acquisition, this approachrequires on-line analysis of the EEG, but has the advantageof allowing activation and control periods to be individuallysampled to maintain the desired ratio of activation to controlimage, which substantially reduces the number of acquisitionperiods required.We used the Stimulate software package because it allowsnon-periodically acquired data to be processed conveniently.As other software packages suitable for non-periodic para-digms (e.g. SPM99) e available, more sophisticatedprocessing of event-related fMRI will be possible.In contrast to the mapping of ictal activity with fMRI,which, due to formidable clinical and technical problems, isrestricted to exceptional cases (e.g. focal status epilepticusor serial seizures without gross movement) (Jackson et al.,1994; Detre et al., 1995), the mapping of interictal activityhas several advantages: (i) frequent interictal discharges mon phenomenon in patients w(ii) interictal discharges are not associated with stimulus- and (iii) fMRI activations associated withsingle discharges are less likely to be confounded withpropagation pared with ongoing ictal activity.By using a method for the on-line subtraction of pulseartefact, we were able to apply our method to patients withlow-amplitude epileptiform discharges. Pulse artefact is themost signicant EEG artefact caused by the ic eldsof the MR scanner, persists throughout the recording and canobscure EEG events with a smaller amplitude (Ives et al.,1993; Huang-Hellinger et al., 1995; Allen et al., 1999). Ithas large inter-individual variability and its occurrence isunpredictable. Frontal and central EEG channels areInterictal EEG-triggered fMRI 1687predominantly affected, with a mean artefact amplitude of50 V in the majority of subjects (Allen et al., 1999). Inthis study, on-line pulse artefact subtraction was essential ineight of 10 patients to reliably detect epileptiform discharges(Fig. 1).Our results suggest that high-amplitude discharges aremore likely to be associated with a larger focal fMRIactivation (e.g. patients 1 and 4). Corresponding to this, twoof the three patients without fMRI activation had particularlow-amplitude discharges with variable morphology (bothspikes and sharp waves). It is not clear, however, why patient7, who had stereotyped discharges with a relatively highamplitude, did not show an fMRI activation and why theactivation in patient 9 was not reproducible. While ourmethod bining EEG and fMRI offers the possibilityof the highly specic detection of local changes in bloodoxygenation associated with interictal epileptiformdischarges, it might not be sensitive enough to detect allactivated areas. Further improvements of the signal-to-noiseratio are therefore required. In particular, the thresholdingapplied may also account for the relatively small size of thefMRI pared with the cortical areas involvedin generating epileptiform discharges found by electro-corticography.Clinical relevanceThe main diagnostic question in patients with localization-related epilepsy, particularly in presurgical evaluation, is tolocalize the area of brain necessary to generate seizures, the‘epileptogenic zone’. fMRI triggered after interictalepileptiform discharges localizes the brain areas involved ingenerating these particular EEG events. The area of cortexthat generates interictal spikes is labelled the ‘irritative zone’.This is not necessarily identical with the cortical area thatinitiates seizures, the ‘ictal onset zone’, but typically has aclose spatial relationship to it (Ebersole and Wade, 1991;Lu¨ders et al., 1996). In studies with patients undergoingepilepsy surgery, the distribution of interictal spikes has beenshown to be a good predictor of surgical e (Gilliamet al., 1997). Hence, the localization of brain areascontributing to the irritative zone by fMRI has the potentialto e a useful additional non-invasive method in thepresurgical evaluation of patients with intractable epilepsy.To determine the signicance of our fMRI ndings, furtherwork is needed to study the results in relation to theanatomical extent of the spiking cortex identied byelectrocorticography and to the surgical e with respectto the extent of removal of the activated area in those patientswho subsequently undergo epilepsy surgery. Furthermore,the distribution of fMRI-derived cortical activation could beused to constrain generator modelling of the scalp-recordedepileptiform discharges, and thereby may be helpful inaddressing the inverse problem, which limits the interpretationof scalp EEG. A recent case report revealed co-localizationof fMRI signal changes triggered by interictal epileptiformactivity and three-dimensional EEG source localization in apatient with multifocal localization-related epilepsy (Seecket al., 1998). These results, however, have to be conrmedin larger groups of patients with stereotyped focal dischargesas presented in our study.In conclusion, EEG-triggered fMRI can reproduciblyvisualize the brain areas involved in generating interictalepileptiform discharges with high spatial resolution. Thisnon-invasive method therefore has the potential to improveour understanding of the pathophysiology of epilepsy andthe interpretation of scalp EEG ndings, and to assist in thepresurgical evaluation of patients with intractable partialseizures.AcknowledgementsThis study was partly funded by the Medical ResearchCouncil, UK. It also received support from the NationalSociety for Epilepsy, Chalfont St Peter, and the Sir JulesThorn Telemetry Unit, National Hospital for Neurology andNeurosurgery, Queen Square, London, UK. K.K. and L.L.are funded by the Medical Research Council, M.R.S. by theBrain Research Trust and G.J.B. by the Multiple SclerosisSociety of Great Britain and Northern Ireland.ReferencesAllen PJ, Polizzi G, Krakow K, Fish DR, Lemieux L. Identicationof EEG events in the MR scanner: the problem of pulse artifactand a method for its subtraction. Neuroimage 9–39.Bullmore E, Brammer M, Williams SC, Rabe-Hesketh S, Janot N,David A, et al. Statistical methods of estimation and inference forfunctional MR image analysis. Magn Reson Med 1–77.Detre JA, Sirven JI, Alsop DC, O’Connor MJ, French JA.Localization of subclinical ictal activity by functional icresonance imaging: correlation with invasive monitoring. AnnNeurol 8–24.Ebersole JS. EEG and MEG dipole source modeling. In: Engel J,Pedley TA, editors. Epilepsy. prehensive textbook.Philadelphia: Lippincott-R 1998. p. 919–35.Ebersole JS, Wade PB. Spike voltage topography identies twotypes of frontotemporal epileptic foci [ments]. Neurology25–1433. Comment in: Neurology 42–3.Engel J Jr, Kuhl DE, Phelps DE, Crandall PH. Comparativelocalization of epileptic foci in partial epilepsy by PCT and EEG.Ann Neurol 9–37.Engel J Jr, Kuhl DE, Phelps ME, Rausch R, Nuwer M. Localcerebral metabolism during partial seizures. Neurology –13.Gilliam F, Bowling S, Bilir E, Thomas J, Faught E, Morawetz R,et al. Association bined MRI, interictal EEG, and ictal EEGresults with e and pathology after temporal lobectomy.Epilepsia 15–20.Hennig J, Janz C, Speck O, Ernst T. Functional spectroscopy of1688 K. Krakow et al.brain activation following a single light impulse: examinations ofthe mechanisms of fast initial response. Int J Imaging SystemsTechnol 3–8.Huang-Hellinger FR, Breiter HC, McCormack G, Cohen MS,Kwong KK, Sutton JP, et al. Simultaneous functional icresonance imaging and electrophysiological recording. Hum BrainMapp –23.Ives JR, Warach S, Schmitt F, Edelman RR, Schomer DL. Monitoringthe patient’s EEG during echo planar MRI. Electroencephalogr ClinNeurophysiol 7–20.Jackson GD, Connelly A, Cross JH, Gordon I, Gadian DG.Functional ic resonance imaging of focal seizures. Neurology0–6.Lee BI, Markand ON, Siddiqui AR, Park HM, Mock B, WellmanHH, et al. Single photon puted tomography (SPECT)brain imaging using N,N,N -trimethyl-N -(2 hydroxy-3-methyl-5–123I-iodobenzyl)-1,3-propanediamine 2 HCl (HIPDM): plex partial seizures. Neurology 71–7.Lemieux L, Allen PJ, Franconi F, Symms MR, Fish DR. Recordingof EEG during fMRI experiments: patient safety. Magn Reson Med3–52.Lu¨ders HO, Engel J Jr, Munari C. General principles. In: Engel JJr, editor. Surgical treatment of the epilepsies. 2nd ed. Philadelphia:Lippincott-R 1996. p. 137–53.Rosen BR, Buckner RL, Dale AM. Event-related functional MRI:past, present, and future. [Review]. Proc Natl Acad Sci USA 3–80.Seeck M, Lazeyras F, Michel CM, Blanke O, Gericke CA, Ives J,et al. Non-invasive epileptic focus localization using EEG-triggeredfunctional MRI and ic tomography.Electroencephalogr Clin Neurophysiol : 508–12.Strupp JP. Stimulate: a GUI based fMRI analysis software package.Neuroimage
Pt 2): S607.Symms MR, Allen PJ, Woermann FG, Polizzi G, Krakow K,Barker GJ, et al. Reproducible localisation of interictal epileptiformdischarges using EEG correlated fMRI. In: Proceedings of theInternational Society for ic Resonance in Medicine, Sydney.1998. p. 168.Warach S, Ives JR, Schlaug G, Patel MR, Darby DG, ThangarajV, et al. EEG-triggered echo-planar functional MRI in epilepsy.Neurology –93.Received September 30, 1998. Revised February 5, 1999.Accepted March 26, 1999播放器加载中，请稍候...
该用户其他文档
下载所得到的文件列表EEG-triggered functional MRI of interictal epileptiform activity in patients with partial seizures.pdf
文档介绍：
EEG-triggered functional MRI of interictal epileptiform activity in patients with partial seizures Brain (1999), 122, EEG-triggered functional MRI of interictalepileptiform activity in patients with partialseizuresK. Krakow,1,4 F. G. Woermann,1,4 M. R. Symms,1,4 P. J. Allen,3 L. Lemieux,1,...
内容来自淘豆网转载请标明出处.}