Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print April 10, 2002, 10.1148/radiol.2233010360
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2233010360v1
223/3/860    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jäger, L.
Right arrow Articles by Reiser, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jäger, L.
Right arrow Articles by Reiser, M.
(Radiology 2002;223:860-869.)
© RSNA, 2002

Technical Developments

Focal Epileptiform Activity in the Brain: Detection with Spike-related Functional MR Imaging—Preliminary Results1

Lorenz Jäger, MD, Konrad J. Werhahn, MD, Alexander Hoffmann, PhD, Sophia Berthold, Verena Scholz, MD, Jürgen Weber, PhD, Soheyl Noachtar, MD and Maximilian Reiser, MD

1 From the Institute of Clinical Radiology (L.J., A.H., S.B., V.S., J.W., M.R.) and Department of Neurology (K.J.W., S.N.), University of Munich, Klinikum Grosshadern, Marchioninistrasse 15, 81366 Munich, Germany. Received January 24, 2001; revision requested March 21; revision received September 7; accepted October 16. Address correspondence to L.J. (e-mail: jaeger@ikra.med.uni-muenchen.de).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and Methods
Results
 Discussion
 REFERENCES
 
The purpose of this study was to develop a spike-related functional magnetic resonance (MR) imaging method to detect epileptic brain activity. Correlations between simultaneous spike-related functional MR imaging and electroencephalographic (EEG) recordings were performed in 10 patients with focal epilepsy. Postprocessing techniques were implemented to eliminate contamination of the EEG recording from ballistocardiography and the echo-planar MR imaging sequence. A diagnostic EEG recording was achieved during functional MR imaging. Spike location correlated with the site of blood oxygen level–dependent signal increase. Spike-related functional MR imaging is a promising technique for detecting focal epileptic brain activity.

© RSNA, 2002

Index terms: Brain, MR, 134.121412 • Electroencephalography (EEG), 13.1299 • Epilepsy • Magnetic resonance (MR), functional imaging


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and Methods
Results
 Discussion
 REFERENCES
 
Determination of the exact localization of epileptic foci is crucial for therapy planning for patients with focal epilepsy. A combination of several techniques is currently used to meet this demand (1), including magnetic resonance (MR) imaging (2), proton MR spectroscopic imaging (3), electroencephalography (EEG) (4) with ictal video monitoring, single-photon emission computed tomography (SPECT) (5,6), positron emission tomography (PET) (7–15), and neuropsychologic imaging. MR imaging, however, can delineate only anatomic abnormalities or lesions such as tumors that cause the epilepsy. In the absence of structural abnormalities, functional imaging modalities are increasingly important. In comparison with the other examination modalities, EEG specifically depicts epileptic abnormalities and has a superior temporal resolution. Major shortcomings of EEG are poor spatial resolution and difficult detection of deep-seated epileptic cortical areas (16).

Results in ictal SPECT studies show a correlation between a single focal area of hyperperfusion and an area of maximum increase in periodic lateralized epileptiform discharges (6,17). However, regional cerebral blood flow may also increase during periodic lateralized epileptiform discharges that consist of unilateral or focal spikes or sharp waves, even when they are not accompanied by a seizure (5,6,11). Major limitations of SPECT are poor spatial and temporal resolutions. The fluorodeoxyglucose (FDG) method is the PET technique used most frequently in the examination of patients with focal epilepsy. After intravenous injection of FDG, the intracerebral uptake and phosphorylation of FDG last 30–40 minutes. Therefore, the temporal and spatial resolutions of FDG PET are limited.

Findings from previous work have shown that acquisition of EEG recordings in the MR environment is a safe procedure if certain precautions are taken (18–26). The quality of the EEG recordings is influenced by the magnetic field, switching gradients, and the radio-frequency pulses. Detection of a proper EEG recording during extended acquisition of functional MR imaging data is not possible, and EEG recordings in MR have been restricted to monitoring between the functional MR imaging data acquisitions (18–21,25,26). This restriction is a major obstacle because baseline data acquisition, which is a period without any epileptiform activity in the EEG recording, is needed to evaluate blood oxygen level–dependent (BOLD) functional MR imaging acquisitions (27–31). With the BOLD technique, however, spike-related functional MR imaging is possible because the radio-frequency pulses are applied with a latency of 3–5 seconds after event detection.

The time delay of BOLD contrast in functional MR imaging is approximately 3 seconds. To exclude the influence on image contrast of any undetectable EEG event during MR, imaging must be started 3–5 seconds after the event and must be finished 3 seconds later. Therefore, data acquisition is limited to the delay time of BOLD contrast after an EEG event (18–21,25), although BOLD contrast data would be available for approximately 10–20 seconds or even longer. To overcome these limitations, we established a method for postprocessing of the EEG recording. Resolution of the EEG recording from interferences due to radio-frequency pulses and gradient switching of a single-shot echo-planar pulse sequence results in a diagnostic EEG recording during functional MR imaging data acquisition.

Cardiac activity causes endogenous microvibrations of the body and micromovements of the head. Micromovements of the head are transmitted from the human scalp to the scalp electrodes to the conducting wires from these electrodes to the amplifier. Motion of the scalp electrodes and the electrode wires in the magnetic field evoke an electric current in these electrodes because they function as an electric loop. This induced current causes a rhythmic excitation with a recurring shape pattern in the EEG recording. We established a method of postprocessing to resolve this ballistocardiographic (32) contamination in the EEG recording. The main objectives of our study were to establish a technique of spike-related functional MR imaging to examine patients with focal epilepsy and to correlate the results with interictal scalp EEG recordings.


   Materials and Methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 REFERENCES
 
Patients
From December 1998 through May 1999, 10 right-handed patients (six women, four men; age range, 19–59 years; mean age, 34 years) with a confirmed diagnosis of focal epilepsy were consecutively examined with functional MR imaging. One patient underwent a second functional MR imaging examination 3 months after the first. The duration of epilepsy ranged from 5 to 40 years (mean, 21 years). The ethical committee of our university approved this study. All patients gave written consent after they received detailed written and oral information about the goals and risks of the functional MR imaging examination.

Before the functional MR imaging examination was performed, all patients underwent high-spatial-resolution MR examinations and then continuous 64-channel scalp EEG monitoring for at least 5 days. Only patients with at least 10 interictal spikes per hour in their previous EEG recording were included in our study. On the basis of long-term EEG monitoring, we included only patients in whom the majority (at least 90%) of the spikes occurred in only one location. The examiners (L.J., A.H., K.J.W.) who performed and evaluated the functional MR imaging examination were unaware of the results of recent clinical examinations (eg, MR imaging, EEG). Patient data are summarized in Table 1.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Clinical Characteristics of Patients

 
EEG Recording in MR Environment
The device required to restore an EEG recorded in an MR environment is an EEG amplifier, which cannot be saturated by any MR activity. With our cooperation, a 32-channel multipurpose electrophysiologic amplifier (EMR; Schwarzer, Munich, Germany) was specifically developed and designed for operation in an MR system. The EEG recordings were performed as recently published (33). Twenty-one sintered silver and silver chloride scalp electrodes were placed on the scalp according to the 10–20 standard electrode positions, with an electrode cap. Electrode-skin impedance was less than 5 k{Omega}. The EEG recording was acquired and displayed by using software (BrainLab; OSG, Rumst, Belgium). A montage to a common reference of several electrodes (Goldman), a referential montage to the electrode position of control zero, or a bipolar longitudinal montage was performed.

Three channels were simultaneously recorded at electrocardiography. Ballistocardiographic contamination of the EEG recording was eliminated by means of postprocessed electrocardiography-triggered subtraction of the averaged ballistocardiogram, according to a previously published method (22,34), with a tool developed in-house in cooperation with a manufacturer (Femr; Schwarzer, Munich, Germany). Another postprocessing tool (24), which was based on commercially available software (Matlab; Math Works, Natick, Mass), was used to eliminate artifacts in the EEG recordings due to performance of the BOLD MR sequences.

To compare the quality of EEG recordings with our set-up, all 10 patients underwent EEG in the supine position both outside the shielded MR room and inside the MR imager before the functional MR imaging experiment. The EEG recordings were again obtained during the functional MR imaging study. The EEG recordings were evaluated online by a neurologist (K.J.W.) with several years of experience in reading EEG recordings, especially those from patients with epilepsy. To ensure the lack of epileptiform discharges during the baseline acquisition of all functional MR imaging studies (spike related and finger tapping), we eliminated all artifacts in the EEG recording that were due to the echo-planar MR imaging sequence by using a previously described postprocessing tool (24). Therefore, sequence artifacts did not mask EEG recordings, and all epileptiform discharges during baseline acquisitions were detected, and corresponding images were excluded from further evaluation (Fig 1).



View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. Patient 3. (a) EEG recording (montage to a common reference of several electrodes) acquired inside the MR imager as a baseline (off state). Artifacts due to the echo-planar MR imaging sequence (short arrow) (repetition time msec/echo time msec of 163,000/64, 5-mm section thickness, 128 x 64 matrix, 210 x 280-mm field of view) disturb the EEG recording (long arrow), which is unreadable. (b) Postprocessed EEG recording of a. All artifacts due to the simultaneous performance of the echo-planar MR imaging sequence are eliminated, and the final EEG recording is of diagnostic quality. During baseline acquisition, spikes (arrowheads) and slow wave complexes (arrows) are clearly detected; therefore, these baseline data were discarded.

 


View larger version (53K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. Patient 3. (a) EEG recording (montage to a common reference of several electrodes) acquired inside the MR imager as a baseline (off state). Artifacts due to the echo-planar MR imaging sequence (short arrow) (repetition time msec/echo time msec of 163,000/64, 5-mm section thickness, 128 x 64 matrix, 210 x 280-mm field of view) disturb the EEG recording (long arrow), which is unreadable. (b) Postprocessed EEG recording of a. All artifacts due to the simultaneous performance of the echo-planar MR imaging sequence are eliminated, and the final EEG recording is of diagnostic quality. During baseline acquisition, spikes (arrowheads) and slow wave complexes (arrows) are clearly detected; therefore, these baseline data were discarded.

 
To verify the spikes that were taken as triggers for functional MR imaging, the EEG recording was postprocessed by eliminating ballistocardiographic contamination of the EEG recording. This allowed clear detection of spikes in the EEG recording. Finally, the postprocessed EEG recording was evaluated a second time by the same neurologist who performed the online EEG reading during the functional MR imaging examination. If the acquisition was started after a false-positive event in the EEG recording, the images were discarded. To image the area of peak activation in the EEG recording on the head surface, amplitude mapping was performed on this postprocessed EEG recording by using the same software. This software was also used to calculate the mean amplitude of the spikes. None of the patients in our functional MR imaging study reported any discomfort or other side effect due to the EEG recording procedure.

Functional MR Imaging Experiment and Postprocessing
The functional MR imaging study was performed with a 1.5-T whole-body MR system (Vision; Siemens, Erlangen, Germany) by using a circular polarized head coil. The MR system was equipped with an echo-planar imaging booster at a rise time of 300 µsec. A sagittal T1-weighted three-dimensional magnetization-prepared rapid gradient-echo sequence (11.4/4.4, field of view of 270 x 270 mm, 256 x 256 matrix, 1.05-mm section thickness, 1.05 x 1.05-mm spatial resolution) was performed to obtain anatomic reference images with a high spatial resolution to allow anatomic fusion with the functional MR images. For BOLD imaging, a single-shot echo-planar pulse sequence (163,000/64, 210 x 280-mm field of view, 128 x 64 matrix, 5-mm section thickness, 2.19 x 3.28-mm spatial resolution) was performed in an oblique-axial projection.

Before the spike-related functional MR imaging study was started for patients 1–4 (Table 1), a functional MR imaging experiment was performed with a standardized finger-tapping paradigm to compare the results of spike-related functional MR imaging with a well-known and established paradigm (27,35). The four patients were asked to perform right- and left-handed finger tapping with the thumb and the middle finger. Four baseline states without finger tapping and four activation states with finger tapping were acquired alternately. Each state comprised eight measurements with 10 sections each. These 10 sections were positioned in the central region. During this finger-tapping functional MR imaging examination, a concurrent EEG recording was obtained and postprocessed to ensure that there were no epileptiform discharges that may have had an effect on the outcome of the study.

The spike-related functional MR imaging examination was carried out in the transverse projection. A series of images with 10 measurements of 10 sections each were collected for the baseline and activation states. These 10 sections covered the suspicious area indicated by the EEG recordings obtained outside the shielded MR room immediately before the functional MR imaging examination. BOLD images were acquired after baseline and activation states, as defined by the neurologist on the basis of online EEG visual inspection. Baseline images were collected after at least 60 seconds without epileptiform activity. The state of activation was defined as a period of time that started with a single spike and lasted as long as 20 seconds. BOLD images were acquired after a delay of approximately 2.5 seconds after detection of a single spike in the EEG recording. Acquisition of baseline and spike-related functional MR images was alternated throughout the examination. The functional MR imaging examination was limited to 1 hour.

The functional MR images were evaluated with software (36,37). All functional data were corrected for three-dimensional motion (38). Activation maps were generated by means of cross correlation (P < .001) (28) between the time series data and an ideal reference function that corresponded to the paradigm timing. To achieve increased T2* and reduced T1 contrasts with a steady state of the spins, the first two measurements of each activation and baseline examination were discarded. Data from corresponding examinations in each subject were averaged. Talairach coordinates were used to determine the location of these areas of activation. The level of significance was calculated for these activated areas, as well as for the signal intensity increase on the BOLD images.

We do not know how many data sets are needed to evaluate spike activity with functional MR imaging. The maximum change of blood oxygenation level detected at functional MR imaging occurs between approximately 4 and 7 seconds after the onset of brain activity (39,40). Therefore, we evaluated the functional MR imaging data between measurements 3 and 5, which were acquired at approximately 6–10 seconds after spike detection in the EEG recording, and between measurements 3 and 7, which were acquired at approximately 6–14 seconds after spike detection. Significant activation was defined as focal activation depicted on at least five images, with a time series of activation concordant with the paradigm of evaluation and reproducible with a level of significance of P < .001. The volume of the activated areas was clustered (5-mm connectivity radius; volume threshold, 500 mm3; single voxel volume, 1 mm3). For each activated area, a time series analysis was calculated with measurements 3–10 to determine the signal intensity increase as a function of time on the BOLD images.

Localization of cortical spike-related functional MR imaging activity was compared with the spike localization in the 64-channel scalp EEG recording and with the amplitude mapping of the EEG recording during the functional MR imaging study. The correlation coefficient, the level of significance, and the signal intensity increase on the BOLD images of the activated areas in the spike-related functional MR imaging studies were correlated with the finger-tapping functional MR imaging studies acquired for patients 1–4. A Spearman rank correlation test was performed to evaluate the correlation between the mean amplitude of the referential montage to control zero in the concordant EEG recordings and the signal intensity increase and the volume of activation on spike-related functional MR images. A Wilcoxon signed rank test was used to evaluate the correlation of findings at functional MR imaging between measurements 3–5 and measurements 5–7.


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
Discussion
 REFERENCES
 
After elimination of ballistocardiographic contamination of the EEG recording (Fig 2) and artifacts due to the echo-planar MR imaging sequence (Fig 1), the EEG recordings acquired inside the MR imager were of diagnostic quality in all 10 patients. Spikes recorded during functional MR imaging had configuration, localization, and amplitude comparable with those recorded outside the MR room immediately before functional MR imaging. In five of 10 patients, functional MR imaging data were not obtained or analyzed. Three patients did not have epileptiform discharges in the 15 minutes of EEG recording performed before the functional MR imaging examination or during the functional MR imaging examination in the MR imager. Therefore, BOLD acquisitions were not obtained in these three patients. In one patient, there were more than 15 spikes per minute throughout the entire functional MR imaging study. Because this patient also had continuous epileptiform activity during all baseline acquisitions that was detected after correction of the EEG recording for pulse artifact, an evaluation of the functional MR imaging data was not possible. One patient was excluded from the evaluation because head motion caused strong motion artifacts in the MR data that could not be corrected. During the BOLD acquisition there were no clinically or EEG-detectable ictal events. The sintered silver and silver chloride scalp electrodes caused only small signal dropouts that were limited to the scalp on the echo-planar and magnetization-prepared rapid gradient-echo images.



View larger version (77K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a. (a) EEG recording (montage to a common reference of several electrodes) acquired inside the MR imager. Spikes (long arrows) are hidden by ballistocardiographic contamination of the EEG recording (small arrowheads). In one case (*), the amplitude of the spike is too small to be differentiated from the ballistocardiographic contamination of the EEG recording. Under the bottom row of the EEG recording, a simultaneous electrocardiographic recording is shown with typical PQRST complex (short arrow). Acquisition of the spike-related BOLD data was started approximately 2.5 seconds after the spike event in the EEG recording, which is indicated by artifacts due to the echo-planar MR imaging sequence (large arrowhead). (b) Fully postprocessed EEG recording of a (montage to a common reference of several electrodes) without ballistocardiographic contamination and without artifacts due to the echo-planar MR imaging sequence. The EEG recording is of diagnostic quality. Single spikes (long arrows) are now clearly detected, including the one (*) that was not clearly detectable in a. Under the bottom row of the EEG recording, the simultaneously acquired electrocardiographic recording shows the PQRST complex of the recorded EEG (short arrow).

 


View larger version (47K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b. (a) EEG recording (montage to a common reference of several electrodes) acquired inside the MR imager. Spikes (long arrows) are hidden by ballistocardiographic contamination of the EEG recording (small arrowheads). In one case (*), the amplitude of the spike is too small to be differentiated from the ballistocardiographic contamination of the EEG recording. Under the bottom row of the EEG recording, a simultaneous electrocardiographic recording is shown with typical PQRST complex (short arrow). Acquisition of the spike-related BOLD data was started approximately 2.5 seconds after the spike event in the EEG recording, which is indicated by artifacts due to the echo-planar MR imaging sequence (large arrowhead). (b) Fully postprocessed EEG recording of a (montage to a common reference of several electrodes) without ballistocardiographic contamination and without artifacts due to the echo-planar MR imaging sequence. The EEG recording is of diagnostic quality. Single spikes (long arrows) are now clearly detected, including the one (*) that was not clearly detectable in a. Under the bottom row of the EEG recording, the simultaneously acquired electrocardiographic recording shows the PQRST complex of the recorded EEG (short arrow).

 
For patients 1–5 (Table 1), the number of spikes during the functional MR imaging examination varied between 10 and 25 (mean, 17 ± 5 [SD]) (Table 2), which determined the number of studies. With use of referential montage to control zero, the mean amplitude of interictal spike activity was in the range of 58–298 µV (mean amplitude of all six functional MR imaging examinations, 119 µV ± 45) (Table 2). Significant focal activation was found in five patients (Table 2) (Fig 3). The localization of cortical activation at functional MR imaging correlated with the activated area at amplitude mapping of the concordant EEG recordings, as well as with the region of activation in the previous routine 64-channel scalp EEG recordings (Tables 1, 2). For patient 3, the functional MR imaging data were repeated in a second EEG-triggered functional MR imaging examination 3 months after the first.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Results of Spike-related Functional MR Imaging

 


View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. Patient 5. (a) Amplitude map of the postprocessed EEG recording. Both electrodes with dark red color (arrow) indicate the location of strong spike activity in the left temporal lobe during spike-related functional MR imaging. (b) Superimposed clustered (P <= .001) BOLD functional MR imaging data on coronal reformatted T1-weighted magnetization-prepared rapid gradient-echo image acquired in the evaluation of measurements 3-7. Colored pixels (arrow) indicate the volume of activation due to spike activity in the left temporal lobe. (c) The time course analysis of the BOLD functional MR imaging data shows the change of signal intensity due to spike activity in the left temporal lobe. This area includes eight voxels in the both the x and y axes. The highest signal intensity increase is found in the center (large solid arrow), which correlates with the ideal reference function (arrowhead) of the paradigm timing. The signal intensity decreases (open arrow) to the periphery of the activated volume. Regions without activation (small solid arrow) have signal intensity at baseline level and do not show any signal intensity change due to inflow or saturation phenomena.

 


View larger version (157K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. Patient 5. (a) Amplitude map of the postprocessed EEG recording. Both electrodes with dark red color (arrow) indicate the location of strong spike activity in the left temporal lobe during spike-related functional MR imaging. (b) Superimposed clustered (P <= .001) BOLD functional MR imaging data on coronal reformatted T1-weighted magnetization-prepared rapid gradient-echo image acquired in the evaluation of measurements 3-7. Colored pixels (arrow) indicate the volume of activation due to spike activity in the left temporal lobe. (c) The time course analysis of the BOLD functional MR imaging data shows the change of signal intensity due to spike activity in the left temporal lobe. This area includes eight voxels in the both the x and y axes. The highest signal intensity increase is found in the center (large solid arrow), which correlates with the ideal reference function (arrowhead) of the paradigm timing. The signal intensity decreases (open arrow) to the periphery of the activated volume. Regions without activation (small solid arrow) have signal intensity at baseline level and do not show any signal intensity change due to inflow or saturation phenomena.

 


View larger version (136K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3c. Patient 5. (a) Amplitude map of the postprocessed EEG recording. Both electrodes with dark red color (arrow) indicate the location of strong spike activity in the left temporal lobe during spike-related functional MR imaging. (b) Superimposed clustered (P <= .001) BOLD functional MR imaging data on coronal reformatted T1-weighted magnetization-prepared rapid gradient-echo image acquired in the evaluation of measurements 3-7. Colored pixels (arrow) indicate the volume of activation due to spike activity in the left temporal lobe. (c) The time course analysis of the BOLD functional MR imaging data shows the change of signal intensity due to spike activity in the left temporal lobe. This area includes eight voxels in the both the x and y axes. The highest signal intensity increase is found in the center (large solid arrow), which correlates with the ideal reference function (arrowhead) of the paradigm timing. The signal intensity decreases (open arrow) to the periphery of the activated volume. Regions without activation (small solid arrow) have signal intensity at baseline level and do not show any signal intensity change due to inflow or saturation phenomena.

 
The location of activation differed by 4–6 mm (x axis, 4 mm; y axis, 6 mm; z axis, 5 mm). In addition to focal activity, patient 3 also showed spike activity (12 spikes in examination 1 and 10 in examination 2) in different frontal areas of both cerebral hemispheres. However, this activity was not used in the evaluation of the functional MR images because it was not reproducible in one location and therefore did not meet the criteria for focal activity. The time course of the signal intensity in each voxel was correlated with an ideal reference function that corresponded to the paradigm timing by using a cross correlation. Voxels that correlated at the level of significance of P < .001 were displayed as active. The mean volume of the clustered activated voxels was 591 mm3 ± 276 for the three measurements between the third and fifth acquisitions and 597 mm3 ± 273 for the five measurements between the third and seventh acquisitions (Table 2).

For the activated voxels, the level of significance was calculated for each subject; it ranged from a P value of 6 x 10-13 to 4 x 10-34 for measurements 3–5 to a P value of 2 x 10-16 to 4 x 10-38 for measurements 3–7 (Table 2). In the activated voxels (P < .001), the mean increase in signal intensity was 15% ± 9, which did not differ significantly for the evaluations between measurements 3 and 5 and between measurements 3 and 7. In all six functional MR imaging studies, the strongest mean signal intensity of activation was achieved for measurement 3, between 6 and 7 seconds after spike detection (Fig 4). The signal intensity of the activation then decreased slightly until measurement 3 was performed, approximately 10 seconds after spike detection. Signal intensity then stayed constant until measurement 8, approximately 14–15 seconds after spike detection. During measurements 9 and 10, about 16–18 seconds after spike detection, signal intensity decreased to nearly baseline level.



View larger version (15K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Time course of mean signal intensity in the volume of activation for all six functional MR imaging examinations. The time course of mean signal intensity due to spike-related activation ({square}) is shown in relation to that during periods of time without spikes (*). Time points for measurements 3-10 are positioned in the middle of each measurement. The strongest increase of signal intensity was found between 6 and 7 seconds after spike detection. After about 18 seconds, signal intensity due to spikes reached nearly baseline level, which represents the state of rest.

 
For the finger-tapping paradigms, the mean increase of signal intensity was 2.99% ± 0.51, and the mean level of significance was P = 3.09 x 10-5 ± 5.14). In all six cases with focal activation depicted on BOLD functional MR images, there was no correlation between the mean amplitude of spikes and the increase of signal intensity at functional MR imaging with referential montage to control zero. However, the mean amplitude of spikes correlated with the volume of activation (P < .025).


   Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
REFERENCES
 
Detection of spike-related local changes in blood oxygenation is based on the statistical evaluation of BOLD functional MR imaging data acquired immediately after an event (on phase) compared with that acquired after a period without any event in the EEG recording (baseline) (18–21). To avoid false-negative baseline acquisitions, which would reduce the total increase of signal intensity on BOLD functional MR images, spike detection in the EEG recording acquired during the echo-planar BOLD sequence is necessary. We implemented a postprocessing procedure (24) to avoid the masking of spikes by echo-planar artifacts and thereby eliminated false-negative baseline data acquisitions. In addition, to improve spike depiction on the MR image, we established a postprocessing method to eliminate ballistocardiographic contamination of the EEG recording (34). Both postprocessing techniques result in diagnostic quality EEG recordings obtained during simultaneous acquisition of echo-planar BOLD images. These processed EEG recordings had the same diagnostic quality as the EEG recordings acquired outside the shielded MR room, which facilitates baseline acquisition by obviating the intravenous injection of clonazepam (19) to reduce spike frequency.

The quality of the echo-planar BOLD MR images was not compromised during EEG recording in the eight patients examined with this technique (one repeated examination for patient 3). Head movement, however, distorted functional MR imaging data so that they were not analyzable despite correction for three-dimensional motion, as was the case in one patient. A further problem with spike-related functional MR imaging is the frequency of spikes. If the frequency is too high, as was the case in one patient, baseline acquisition is not possible and therefore the BOLD technique cannot be applied. Because of these shortcomings, such patients should be excluded from spike-related BOLD functional MR imaging studies.

In all five patients in the current study (especially patient 3, who underwent a second functional MR imaging examination 3 months after the first), the simultaneous EEG-functional MR imaging application resulted in a significant focal signal increase at functional MR imaging that was associated with spike detection. For patient 3, not only was there reproducibility of the signal intensity increase on spike-related BOLD functional MR images but also reproducibility of the localization. This result is in good agreement with findings in BOLD functional MR imaging studies of visual activation patterns (41) and with those in EEG dipole localization of epileptiform activities in EEG recordings (4). The mean amplitude of spikes in the EEG recording during functional MR imaging has proven to be a marker for the total activated volume on spike-related BOLD functional MR images because there was a significant correlation between the total activated volume in the clustered data and the amplitude of spikes in the EEG recording.

In the six cases evaluated, time series analysis showed a peak of signal intensity increase 6–7 seconds after the detected spike; this finding is consistent with results of event-related functional MR imaging with visual stimulation (42). A plateau of high signal intensity appeared between 10 and 15 seconds. However, the cross-correlation coefficient was higher for measurements 3–7 than that for measurements 3–5, with the level of significance higher for the former. The larger data set, which resulted in additional reduction of background noise because of averaging, is the most likely reason for the higher degree of significance in the evaluation of five measurements (measurements 3–7). However, three measurements (measurements 3–5) are sufficient for detection of spike-related focal brain activity. In addition, a mean of only 17 spikes per patient was sufficient to help identification of focal brain activity due to spikes with the echo-planar BOLD technique.

This finding is in contrast to those of others (18,19), who reported that at least 30 spikes are needed for a functional MR imaging examination. The necessity for a larger number of spikes at spike-related functional MR imaging in those studies may have two explanations. First, in those cases with a total acquisition time of 3 seconds or more, the lack of spike detection during baseline acquisition owing to artifacts from the echo-planar MR imaging sequence may result in false-negative baseline data sets. This leads to a low mean difference between baseline data and that acquired in the activated state, which reduces the mean increase of signal intensity on BOLD functional MR images. Second, and even more important, the inability to detect spikes during baseline acquisition (17–20) limits the total acquisition time to only 3 seconds to avoid the influence of false-positive baseline data on the evaluation of the BOLD functional MR imaging data.

Therefore, functional MR imaging measurements were restricted to one or two measurements that each lasted 2–3 seconds. However, magnetization reaches equilibrium during the acquisition of the first two measurements (28,43). This results in a continuous decrease of signal intensity on the BOLD functional MR images. Nevertheless, compared with the steady state, the signal intensity on the images with these first two baseline measurements is still high. If nonrepeated or repeated multisection examinations are performed that are limited to only two measurements, the signal intensity increase and the level of significance will therefore be low on the BOLD functional MR images. Therefore, a total number of at least 30 examinations were needed, which resulted in a minimum of 60 measurements. In contrast to these reports, we successfully performed spike-related BOLD functional MR imaging in 10–25 examinations. However, because we evaluated three to five measurements in each examination, a total of 30–125 measurements were obtained during each functional MR imaging examination. Because of the larger amount of data acquired in our examinations, the areas of activation on BOLD functional MR images are also more significant.

Authors of EEG-related BOLD functional MR imaging studies (18) have reported a spike-dependent signal intensity increase of only 1%–2%, which is half the increase we observed in our finger-tapping study. These findings are different from our results of spike-related functional MR imaging, which showed a mean increase of signal intensity of 15%. In our study, the two smallest increases of signal intensity (7% and 3%) were found for patient 3, who underwent two functional MR imaging examinations performed 3 months apart. In this patient, only 50% of all spikes were located in the same anatomic area, while the other 50% were scattered in both frontal lobes. It should be noted, however, that the exclusion of patient 3 from data analysis still results in a wide range of signal intensity increase for this population. This finding is consistent with those of others, who also reported a wide range of brain activity due to epileptiform discharges in EEG recordings acquired during BOLD functional MR imaging (18, 21,44) and SPECT and PET (5,6,10–12). PET studies with nitrogen 13 hydrogen 3 have revealed that approximately 30% of patients with focal epilepsy showed normal regional cerebral blood flow during the interictal state (12), while SPECT studies showed increased regional cerebral blood flow in the epileptic foci during the interictal state (5). In addition, for patients with focal epilepsy, spikes were found in the EEG recordings that correlated with the signal intensity increase on BOLD functional MR images and with spikes that did not correlate with signal intensity increase (18).

These findings, as well as our own, argue for a wide range of signal intensity increases on spike-related BOLD functional MR images that may be dependent on the heterogeneity of the spike population and on an uncoupling of regional cerebral blood flow and local cerebral metabolic rate of oxygen. This uncoupling was found in SPECT studies (45), with an increase of regional cerebral blood flow, while the local cerebral metabolic rate of oxygen decreased during periodic lateralized epileptiform discharges (44). These hypotheses are supported by the fact that, for each patient, there was not much difference in signal intensity increase on BOLD functional MR images acquired during the finger-tapping experiments. Furthermore, the signal intensity on baseline acquisitions remained constant throughout the BOLD functional MR imaging study.

EEG-triggered spike-related BOLD functional MR imaging, with postprocessed EEG recordings without ballistocardiographic contamination and artifacts due to fast gradient switching by the echo-planar MR imaging sequence, seems to be a promising method for detecting interictal epileptiform activity even if the total number of spikes is rather small. The combination of EEG and spike-related BOLD functional MR imaging brings together the advantages of the superior temporal resolution of EEG recording and the high spatial and good temporal resolutions of functional MR imaging.


   ACKNOWLEDGMENTS
 
We thank Laurie Gauger and Dana Selover for reading the manuscript.


   FOOTNOTES
 
Abbreviations: BOLD = blood oxygen level–dependent, EEG = electroencephalography, FDG = fluorodeoxyglucose

Author contributions: Guarantors of integrity of entire study, L.J., S.N., M.R.; study concepts, L.J., K.J.W., A.H.; study design,, L.J., A.H., V.S., J.W.; literature research, L.J., S.B.; clinical studies, V.S., S.N.; experimental studies, L.J., A.H., V.S., K.J.W.; data acquisition, L.J., A.H., K.J.W.; data analysis/interpretation, L.J., S.N., S.B., K.J.W., A.H., J.W.; statistical analysis, L.J., A.H., S.B., J.W.; manuscript preparation, L.J., A.H., S.B.; manuscript definition of intellectual content, L.J., A.H., K.J.W.; manuscript editing, L.J., A.H.; manuscript revision/review and final version approval, S.N., M.R.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 REFERENCES
 

  1. Noachtar S, Lüders HO, Bromfield EB. Surgical therapy of epilepsy. In: Brandt T, Caplan LR, Dichgans J, Diener HC, Kennard C, eds. Neurological disorders: course and treatment. San Diego, Calif: Academic Press, 1996; 183-191.
  2. Otsubo H, Hwang PA, Jay V, et al. Focal cortical dysplasia in children with localization-related epilepsy: EEG, MRI, and SPECT findings. Pediatr Neurol 1993; 9:101-107.[CrossRef][Medline]
  3. Capizzano AA, Vermathen P, Laxer KD, et al. Temporal lobe epilepsy: qualitative reading of 1H MR spectroscopic images for presurgical evaluation. Radiology 2001; 218:144-151.[Abstract/Free Full Text]
  4. Krings T, Chiappa KH, Cuffin BN, Buchbinder BR, Cosgrove GR. Accuracy of electroencephalographic dipole localization of epileptiform activities associated with focal brain lesions. Ann Neurol 1998; 44:76-86.[CrossRef][Medline]
  5. Lee BI, Markand ON, Siddiqui AR, et al. Single photon emission computed tomography (SPECT) brain imaging using N,N,N’-trimethyl-N’-(2 hydroxy-3-methyl-5-123I-iodobenzyl)-1,3-propanediamine 2 HCI (HIPDM): intractable complex partial seizures. Neurology 1986; 36:1471-1477.[Abstract/Free Full Text]
  6. Lee BI, Markand ON, Wellman HN, et al. HIPDM-SPECT in patients with medically intractable complex partial seizures: ictal study. Arch Neurol 1988; 45:397-402.[Abstract]
  7. Barrington SF, Koutroumanidis M, Agathonikou A, et al. Clinical value of "ictal" FDG-positron emission tomography and the routine use of simultaneous scalp EEG studies in patients with intractable partial epilepsies. Epilepsia 1998; 39:759.
  8. Arnold S, Noachtar S, Linke R, et al. Ictal SPECT hyperperfusion reflects the activation of the symptomatogenic cortex in spontaneous and electrically-induced non-habitual focal epileptic seizures: correlation with subdural EEG recordings. Epileptic Disord 2000; 2:41-44.[Medline]
  9. Noachtar S, Arnold S, Yousry TA, Bartenstein P, Werhahn KJ, Tatsch K. Ictal technetium-99m ethyl cysteinate dimer single-photon emission tomographic findings and propagation of epileptic seizure activity in patients with extratemporal epilepsies. Eur J Nucl Med 1998; 25:166-172.[CrossRef][Medline]
  10. Engel J, Kuhl DE, Phleps ME, Crandall PH. Comparative localization of epileptic foci in partial epilepsy by PCT and EEG. Ann Neurol 1982; 12:529-537.[CrossRef][Medline]
  11. Franck G, Sadzot B, Salmon E, et al. Regional cerebral blood flow and metabolic rates in human focal epilepsy and status epilepticus. Adv Neurol 1986; 44:935-948.[Medline]
  12. Kuhl DE, Engel J, Phelps ME, Selin C. Epileptic patterns of local cerebral metabolism and perfusion in humans determined by emission computed tomography of 18FDG and 13NH3. Ann Neurol 1980; 8:348-360.[CrossRef][Medline]
  13. Rowe CC, Berkovic SF, Fracp MD, et al. Patterns of postictal cerebral blood flow in temporal lobe epilepsy: qualitative and quantitative analysis. Neurology 1991; 41:1096-1103.[Abstract/Free Full Text]
  14. Arnold S, Berthele A, Drzezga A, et al. Reduction of benzodiazepine receptor binding is related to the seizure onset zone in extratemporal focal cortical dysplasia. Epilepsia 2000; 41:818-824.[CrossRef][Medline]
  15. Drzezga A, Arnold S, Minoshima S, et al. 18F-FDG PET studies in patients with extratemporal and temporal epilepsy: evaluation of an observer-independent analysis. J Nucl Med 1999; 40:737-746.[Abstract/Free Full Text]
  16. Eldinger G, Wach P, Pfurtscheller G. On the realization of an analytic high-resolution EEG. IEEE Trans Biomed Eng 1998; 45:736-745.[CrossRef][Medline]
  17. Lee BI, Markand ON, Wellman HN, et al. HIPDM single photon emission computed tomography brain imaging in partial onset secondarily generalized tonic-clonic seizures. Epilepsia 1987; 28:305-311.[Medline]
  18. Krakow K, Woermann FG, Symms MR, et al. EEG-triggered functional MRI of interictal epileptiform activity in patients with partial seizures. Brain 1999; 122:1679-1688.[Abstract/Free Full Text]
  19. Lazeyras F, Blanke O, Perring S, et al. EEG-triggered functional MRI in patients with pharmacoresistant epilepsy. J Magn Reson Imaging 2000; 12:177-185.[CrossRef][Medline]
  20. Patel MR, Blum A, Pearlman JD, et al. Echo-planar functional MR imaging of epilepsy with concurrent EEG monitoring. AJNR Am J Neuroradiol 1999; 20:1916-1919.[Abstract/Free Full Text]
  21. Warach S, Ives JR, Schlaug , et al. EEG-triggered echo-planar functional MRI in epilepsy. Neurology 1996; 47:89-93.
  22. Allen PJ, Polizzi G, Krakow K, Fish DR, Lemieux L. Identification of EEG events in the MR scanner: the problem of pulse artifacts and a method for its subtraction. Neuroimage 1998; 8:229-239.[CrossRef][Medline]
  23. Hoffmann A, Faber SC, Werhahn KJ, Jäger L, Reiser M. Electromyography in MRI: first recordings of peripheral nerve activation caused by fast magnetic field gradients. Magn Reson Med 2000; 43:534-539.[CrossRef][Medline]
  24. Hoffmann A, Jäger L, Werhahn KJ, Jaschke M, Noachter S, Reiser M. Electroencephalography during functional echo-planar imaging: detection of epileptic spikes using post-processing methods. Magn Reson Med 2000; 44:791-798.[CrossRef][Medline]
  25. Ives JR, Warach S, Schmitt F, Edelman RR, Schomer DL. Monitoring the patient’s EEG during echo planar MRI. Electroencephalogr Clin Neurophysiol 1993; 87:417-420.[CrossRef][Medline]
  26. Lemieux L, Allen PJ, Franconi F, Symms MR, Fish DR. Recording of EEG during fMRI experiments: patient safety. Magn Reson Med 1997; 38:943-952.[Medline]
  27. Bandenttini PA, Wong EC, Hinks RS, Tikofsky RS, Hyde J. Time course EPI of human brain function during task activation. Magn Reson Med 1992; 25:390-397.[Medline]
  28. Bandettini PA, Jesmanowicz A, Wong EC, Hyde JS. Processing strategies for time course data sets in functional MRI of the human brain. Magn Reson Med 1993; 30:161-173.[Medline]
  29. Kwong KK, Belliveau JW, Chesler DA, et al. Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc Natl Acad Sci U S A 1992; 89:5675-5679.[Abstract/Free Full Text]
  30. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci U S A 1990; 87:9868-9872.[Abstract/Free Full Text]
  31. Ogawa S, Tank D, Menon R, et al. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci U S A 1992; 89:5951-5955.[Abstract/Free Full Text]
  32. Sturm R, Nigg B, Koller EA. The impact of cardiac activity on triaxially recorded endogenous microvibrations of the body. Eur J Appl Physiol Occup Physiol 1980; 44:83-96.[CrossRef][Medline]
  33. Homan RW, Herman J, Purdy P. Cerebral location of international 10–20 system electrode placement. Electroencephalogr Clin Neurophysiol 1987; 66:376-382.[CrossRef][Medline]
  34. Jäger L, Hoffmann A, Joppich M, Reiser M. Simultaneous EEG-recording with MR data-acquisition (abstr) In: Proceedings of the Sixth Meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1998; 286.
  35. Yousry TA, Schmid UD, Jassoy AG, et al. Topography of the cortical motor hand area: prospective study with functional MR imaging and direct motor mapping at surgery. Radiology 1995; 195:23-29.[Abstract/Free Full Text]
  36. Cox RW. Software for analysis and visualization of functional magnetic resonance neuroimages. Comput Biomed Res 1996; 29:162-173.[CrossRef][Medline]
  37. Cox RW, Hyde JS. Software tools for analysis and visualization of FMRI Data. NMR Biomed 1997; 10:171-178.[CrossRef][Medline]
  38. Cox RW, Jesmanowicz A. Real-time 3D image registration for functional MRI. Magn Reson Med 1999; 42:1014-1018.[CrossRef][Medline]
  39. Bandettini PA, Cox RW. Event-related fMRI contrast when using constant interstimulus interval: theory and experiment. Magn Reson Med 2000; 43:540-548.[CrossRef][Medline]
  40. Hennig J, Janz C, Speck O, Ernst T. Functional spectroscopy of brain activation following a single light impulse: examinations of the mechanisms of fast initial response. Int J Imaging Systems Technol 1995; 6:203-208.
  41. Rombouts SARB, Barkof F, Hoogenraad FGC, Sprenger M, Scheltens P. Within-subject reproducibility of visual activation patterns with functional magnetic resonance imaging using multislice echo planar imaging. Magn Reson Imaging 1998; 16:105-113.[CrossRef][Medline]
  42. Rosen BR, Buckner RL, Dale AM. Event-related functional MRI: past, present, and future. Proc Natl Acad Sci U S A 1998; 95:773-780.[Abstract/Free Full Text]
  43. Duyn JH. Steady state effects in fast gradient echo magnetic resonance imaging. Magn Reson Med 1997; 37:559-568.[Medline]
  44. Jackson GD, Connelly A, Cross JH, Gordon I, Gadian DG. Functional magnetic resonance imaging of focal seizures. Neurology 1994; 44:850-856.[Abstract/Free Full Text]
  45. Lee BI, Schauwecker DS. Regional cerebral perfusion in PLEDs: a case report. Epilepsia 1988; 29:607-611.[Medline]



This article has been cited by other articles:


Home page
 BrainHome page
Y. Aghakhani, A. P. Bagshaw, C. G. Benar, C. Hawco, F. Andermann, F. Dubeau, and J. Gotman
fMRI activation during spike and wave discharges in idiopathic generalized epilepsy
Brain, May 1, 2004; 127(5): 1127 - 1144.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
L. Lemieux, A. Salek-Haddadi, K. Krakow, L. Jager, A. Hoffmann, and M. Reiser
The Nature of MR Signal Changes [letter] *Dr Jager and colleagues respond:
Radiology, March 1, 2003; 226(3): 922 - 925.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2233010360v1
223/3/860    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jäger, L.
Right arrow Articles by Reiser, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jäger, L.
Right arrow Articles by Reiser, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE