Brain, Vol. 122, No. 10, 1851-1865,
October 1999
© 1999 Oxford University Press
An H215O-PET study of cerebral blood flow changes during focal epileptic discharges induced by intracerebral electrical stimulation
1 Department of Neurosciences and 2 INSERM 318 Research Unit, Grenoble Hospital, 3 CERMEP and 4 Functional Neurology and Epilepsy Department, Neurological Hospital, Lyon, France
Correspondence to:
Dr P. Kahane, Neurophysiopathologie de l'Epilepsie, Clinique Neurologique, CHU de Grenoble, BP 217 X, 38043 Grenoble cedex, France E-mail: philippe.kahane{at}ujf-grenoble.fr
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsDiscussionReferences |
|---|
Partial epileptic seizures are known to cause a focal increase in cerebral blood flow (CBF). However, quantified studies of ictal CBF changes under intracranial EEG control are still needed to assess the relationships in time and space between CBF changes and electrical discharges. Ten patients undergoing an intracerebral stereotaxic EEG (stereo-EEG) investigation for epilepsy surgery were prospectively studied for local perfusion changes. These were measured by H215O-PET during 12 subclinical or mild symptomatic focal epileptic discharges induced by intracerebral electrical stimulation of the hippocampus (eight), amygdala (two), temporal pole (one) and fusiform gyrus (one). This study aimed to assess whether a significant focal blood flow change reflected the geographical extent of the underlying coincident epileptic discharge, as measured by this method at seizure onset. No significant CBF change was observed on test–retest at rest or during ineffective electrical stimulations outside the epileptogenic area. Compared with the resting condition, a significant focal perfusion increase of 16–55% occurred during eight discharges, there was no CBF change in three and a significant CBF decrease in one. Ictal CBF increases were mostly associated with low-voltage fast activity, but their magnitude had no obvious link with the duration of the discharge (range 8–106 s). Regional analysis of ictal PET was performed in 10 anatomical areas during each of the 12 discharges. Of all the 120 regions, 59 were not explored by intracerebral electrodes and 14 (24%) of these demonstrated ictal CBF changes. In 43 of the 61 regions explored by stereo-EEG (70.5%), PET and depth EEG findings converged, showing either a CBF change in a discharging area or no CBF change in a region unaffected by the discharge. Areas of increased CBF indicated an underlying epileptic discharge in almost 100% of the cases. Conversely, of the 18 regions showing discrepancies between intracerebral recordings and PET data, 17 were discharging regions showing no ictal CBF changes. Thus, a focal CBF increase, when detected at the seizure onset concomitantly with the initial low-voltage fast activity, was a reliable marker of an underlying epileptic discharge. It emphasizes the importance of injecting blood-flow tracers as soon as possible after detection of the discharge in routine clinical studies, even at a subclinical stage of the seizure. However, the extent of significant ictal CBF changes can be more restricted than that of the electrical discharge, thus limiting the reliability of ictal CBF images for outlining the contours of a tailored cortectomy.
cerebral blood flow; partial epileptic seizures; PET; intracerebral recordings; intracerebral electrical stimulation
CBF = cerebral blood flow; SPECT = single photon emission computed tomography; stereo-EEG = intracerebral stereotaxic EEG
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
For over 60 years it has been known that an increase in cerebral blood flow (CBF) can occur during the course of epileptic seizure (Gibbs, 1933
; Penfield, 1933; Gibbs et al., 1934), beginning shortly after the onset of chemically or electrically induced focal cortical discharges (Penfield et al., 1939; Jasper and Erickson, 1941) and preceding changes in pH (Jasper and Erickson, 1941). More recently, observations of ictal focal hyperperfusion have been extended to larger series of patients, thanks to the development of CBF functional imaging techniques (Ingvar, 1973; Kuhl et al., 1980; Franck et al., 1986; Henry, 1996; Theodore et al., 1996; Duncan, 1997a, b). Notably, ictal single photon emission computed tomography (SPECT) has proved to be a reliable marker of the side and, to a lesser degree, of the site and extent of the seizure discharge, at least in temporal lobe epilepsy (Duncan et al., 1993; Grünwald et al., 1994; Spencer, 1994; Newton et al., 1995; Ho et al., 1996).
However, only a few ictal CBF studies have been conducted under intracranial EEG monitoring, using CBF probe monitoring (Dymond and Crandall, 1976; Weinand et al., 1994; Wallstedt et al., 1995), xenon-CT (Johnson et al., 1993) or SPECT (Spanaki-Varelas et al., 1997). All these studies confirmed that focal hyperaemia occurred in direct relation to seizure activity, but most had poor spatial sampling or were unable to assess the statistical significance of ictal CBF changes compared with the interictal state.
To test the hypothesis that a significant focal CBF change detected as soon as possible after the seizure onset reflects the geographical extent of the underlying epileptic discharge, we prospectively studied the local perfusion changes measured by H215O-PET during focal epileptic discharges induced by intracerebral electrical stimulation in 10 patients undergoing an intracerebral stereotaxic EEG (stereo-EEG) investigation. This enabled us to assess (i) the type and amplitude of perfusion changes that can be detected early in the course of focal discharges; (ii) the accuracy with which these changes reflect intracerebral discharges in space and time; and (iii) the relationship between ictal CBF changes and the electrical pattern of discharge.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Selection criteria
For this study, 10 adult patients were selected consecutively among a larger group of 79 epileptic patients who had undergone stereo-EEG at the Grenoble hospital between December 1, 1993 and May 3, 1996. Enrolment in this H215O-PET study under stereo-EEG control was proposed at the end of the investigation to patients in whom intracerebral electrical stimulation had elicited epileptic discharges associated with no or only subjective ictal symptoms, as performed in our routine procedure (see below). These patients were thus likely to lie motionless during ictal PET data acquisition.
All patients entered the study after having given their informed consent according to the recommendations of the ethics committee (CCPPRB) of the Grenoble hospital.
The PET study was performed at the Cyclotron PET facility at the Lyon Neurological Hospital (CERMEP).
Subjects
The 10 patients (five males and five females aged between 23 and 41 years) were all suffering from long-lasting (5–22 years) severe drug-resistant partial epilepsy with frequent seizures (from two or three per week to >30 per month), and all were candidates for epilepsy surgery. Relevant MRI, clinical, stereo-EEG and surgical data are given in Table 1.
|
MRI performed on a 1.5 Tesla device, including 3D T1 acquisition and T2 sequences perpendicular and parallel to the hippocampal plane, was normal in four cases and exhibited signs of mesiotemporal sclerosis in five cases. In the remaining patient, non-specific bilateral parieto-occipital increases in signal were observed in T2-weighted sequences.
All patients first underwent scalp video-EEG monitoring of their spontaneous seizures, on the basis of which a stereo-EEG study was considered necessary to determine the extent of the cortical excision required. Stereo-EEG was performed following our routine procedure (Munari et al., 1994), according to the method developed by Talairach and Bancaud (Bancaud et al., 1965; Talairach et al., 1974; Munari and Bancaud, 1987). The cortical stereo-EEG targets were determined in each patient on the basis of available clinical, scalp EEG and MRI data, in order to assess the origin and early spreading of ictal discharges. Consequently, the number and positions of implanted electrodes varied from one patient to another, including 7–13 intracerebral electrodes of which 52% proved to be located within the epileptogenic area. Electrodes were implanted under stereotaxic conditions (Talairach and Bancaud, 1973) with the help of a computer-driven robot (Benabid et al., 1992). Long-term video stereo-EEG recordings were made over a period of 1–3 weeks, enabling us to record spontaneous seizures and to perform low-frequency (1 Hz) and high-frequency (50 Hz) electrical stimulation according to our routine protocol (Kahane et al., 1993). The aims of such stimulation were to map functionally eloquent areas and to reproduce some or all of the electroclinical seizures.
The origin of seizures proved to be temporal in seven patients, temporo-occipitoparietal (including the speech area) in two and temporofrontal in one. Operations were carried out on all patients and the postoperative follow-up period ranged from 16 to 45 months. Seven of the eight patients in whom removal of the epileptogenic zone was considered complete are totally seizure-free, whereas the two patients in whom the cortectomy could not include the whole epileptogenic area because of anatomical constraints show persisting seizures (Table 1
).
Study design
PET study with simultaneous stereo-EEG recording was performed just before the electrodes were removed after collection of the data necessary for surgical planning. Consequently, the sites and parameters of ineffective and effective stimulation and the types and timing and associated symptoms of electrically induced discharges were defined precisely for each patient before entry into the study.
The study paradigm included six runs of H215O-PET acquisition carried out in a single session under stereo-EEG monitoring (Fig. 1).
|
The first two sets of PET images were acquired without stimulation, to obtain test–retest measurements of CBF at rest. Then, during acquisitions of PET scans 3 and 4, stimulation was applied outside the epileptogenic area. PET data acquired during these ineffective stimulations were used to assess whether the stimulation alone could induce CBF modifications. Two stimulations were delivered in the epileptogenic area during PET scans 5 and 6 for ictal CBF studies. Stimulation parameters used in the latter two sessions were known to be capable of eliciting subclinical or mild symptomatic discharges compatible with PET data acquisition in each individual. In eight patients, electrical stimulation at both 1 and 50 Hz was used; in the remaining two patients only 50-Hz stimulation was applied.
The interval between H215O injection and the onset of stimulation was calculated in order to assess the CBF changes as early as possible with respect to the expected discharge. Thus, the tracer was injected 10–15 s before the onset of the 50-Hz stimulation, since it was known that discharges occurred within the first 5 s following the onset of this type of stimulation. Conversely, for 1-Hz stimulation, this delay was known to vary from 20 to 30 s according to individuals. Consequently, the H215O bolus was injected 5–10 s after stimulation onset.
Stereo-EEG recordings and stimulations
We used semi-rigid intracerebral electrodes 0.8 mm in diameter (Dixi, Besancion, France), featuring 5, 10 or 15 contacts 2 mm long and 1.5 mm apart, depending on the size and depth of the target region.
Stereo-EEG recordings were obtained using bipolar derivations between contiguous contacts, at different levels along the axis of each electrode. Thus, various mesial, basal and lateral structures of anatomical lobes were evaluated, often at different sites within the same structure. It should be noted that the temporal pole was explored in all patients whose seizures arose mostly from the mesiotemporal region. An overview of the cortical areas explored by the electrodes is given in Fig. 6 (see also Results).
|
Each stimulation was performed after checking that depth EEG exhibited its usual baseline activity. Electrical stimuli were delivered using a conventional rectangular pulse generator (World Precision Instruments, New Haven, Conn., USA). They were applied between two contiguous contacts with alternating polarity. Effective and ineffective stimulations were matched in frequency, pulse width and intensity (Fig. 1
).
All recording and stimulation sites, as well as the locations of the induced discharges, were anatomically localized according to their stereotaxic coordinates using Talairach and Tournoux's proportional atlas (Talairach and Tournoux, 1988).
Baseline depth EEG activity was analysed and compared visually during the two resting conditions.
The interval between the tracer injection and the first relevant electrical change and the duration of the discharge were measured.
The electrical patterns recorded during the entire duration of induced discharges were carefully examined in all of the structures involved, and each pattern was classified in one of three categories: (i) low voltage, fast activity; (ii) spike or polyspike discharges of high amplitude; (iii) rhythmic activity (1–5 Hz) of spikes, spikes and waves, polyspikes and waves, or sharp waves.
We checked that no other electrical modification occurred in areas unaffected by the induced discharges, notably for spreading depression.
PET data acquisition and modelling
PET acquisitions
All acquisitions were performed in patients with eyes closed and ears unplugged. Patients were asked to remain silent and motionless if they felt a seizure coming. A description of the symptoms was obtained just after termination of data acquisition.
Acquisitions were performed with a time-of-flight PET scanner (TTV03, LETI; CENG Grenoble, France) generating seven 9-mm-thick slices separated by 3-mm gaps.
The patient was positioned in the scanner so that the slices were parallel to the bicommissural plane (AC–PC) according to external landmarks. Images were reconstructed with a Hanning filter providing a spatial resolution of 7 mm (Trébossen and Mazoyer, 1991). Attenuation correction was performed using measured coefficients derived from a 20 min transmission scan. Images were reconstructed over 90 s, the starting point being the time when activity reached 20% of its peak value. Acquisitions of 90 s were used to maximize the signal-to-noise ratio and to improve individual data analysis. All scans were performed at 20-min intervals, corresponding to 10 half-lives of the tracer. As no arterial catheter was used, the reconstructed images were not converted to regional CBF. However, blood flow over the range tested has been shown to be linearly related to the observed activity (Herscovitch et al., 1983). Therefore, responses reported here are changes in activity distribution but will be referred to as changes in blood flow.
PET data analysis
Activity images were first tested for any movement during or between conditions and were realigned when necessary (Woods et al., 1992). In order to eliminate the influence of global blood flow changes and to focus on regional blood flow modifications, images were normalized proportionally to the global activity of the scan under consideration. Difference images were calculated by subtracting normalized images obtained at rest from stimulation images, and were then analysed as described below. Images at rest were tested against each other previously in order to reveal possible significant changes during non-stimulated interictal scans.
Local changes in CBF between conditions were detected by analysis of individual brain activation maps (Poline and Mazoyer, 1994). For each normalized difference image, the software provided a map of all the clusters of contiguous pixels for which the hypothesis of a pure noise image could be rejected at a given significance level.
Detection of clusters.
The algorithm was based on a hierarchical multiscale description of the PET difference image in terms of connected objects. The multiscale approach refers to the bandwidth of four different Gaussian filters applied successively to the original image. The idea in this procedure was to increase the signal-to-noise ratio of signals of different size, and therefore to improve the overall detection sensitivity. After each filtering step, the image was hierarchically decomposed and clusters were sorted by size and magnitude.
Test
.
For each cluster, parameters (size and amplitude) were simultaneously tested against the hypothesis of pure noise images, the corresponding 2D frequency distribution having been previously assessed by Monte Carlo simulation. In this study, the significance level was set at P < 0.01.
The anatomical localization of significant CBF changes was identified by superimposing maps of significant CBF variations on patients' T1 MRI images after 3D realignment to the PET images (Woods et al., 1993). To calculate the percentage of significant blood flow increase or decrease, regions of interest were drawn following the contours of the map and were reported on the corresponding slice of the reference and activation scans. The mean normalized activity and the standard deviation were obtained for each measure. The magnitude of the CBF change was calculated as a percentage using the following formula:
 |
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
All 20 CBF measurements obtained at rest were valid, as were those carried out under stimulation applied outside the epileptogenic area. Because of the variability in effectiveness of stimulations applied in the epileptogenic area, 14 epileptic discharges were induced in nine of the 10 patients studied instead of the 20 discharges expected. In addition, early movements developed during two of the 14 induced discharges, and the corresponding CBF data were withdrawn from analysis. Thus, 12 `ictal' PET studies were available in nine patients, three of whom showed two elicited discharges. Although PET data acquired during stimulation in the epileptogenic area that failed to trigger any discharge were excluded from the final analysis, we checked that no significant CBF changes occurred in this condition compared with the resting condition.
Test–retest CBF study in resting conditions
No significant test–retest CBF changes were observed at rest. Therefore, when interictal stereo-EEG traces showed similar activities at rest on visual analysis, rest PET scans were considered to be comparable and rest measures were averaged for comparison with stimulation conditions. When interictal stereo-EEG activities were different in the two resting conditions, we used the data obtained in the resting condition during which the activity matched more closely that recorded at stimulation onset. In all patients, we checked that CBF changes between stimulation and rest conditions were stable in space and of similar magnitude when either of the two resting conditions was used as the baseline. We also checked that no spontaneous discharge occurred during or between PET data acquisitions.
CBF study during ineffective electrical stimulations
Low- and high-frequency ineffective stimulations were applied outside the epileptogenic area in various structures, including the frontobasal, frontopolar and mesial frontal cortices, the frontoparietal operculum, the parietal cortex, the fusiform gyrus, the first and second temporal convolutions and the anterior and posterior cingulate gyrus. None of them elicited any clinical or stereo-EEG changes or any significant CBF changes compared with the resting condition.
CBF study during electrically induced focal epileptic discharges
Twelve epileptic discharges were induced by stimulation at 50 Hz (n = 8) or 1 Hz (n = 4), of which nine were symptomatic (Table 2). Stimulation was applied over the right (n = 7) or left (n = 5) hemisphere, within mesial and polar temporal lobe structures in 11 cases and within the fusiform gyrus in the remaining one. The tracer was injected from 3 to 30 s before the discharge onset; the duration of the stereo-EEG discharges varied from 8 to 106 s.
|
Global results
A marked and significant increase in focal CBF was observed during eight discharges (66.5%), ranging from 16 to 55% (Fig. 2
) when compared with the resting condition. Only one of these discharges demonstrated a mixed ictal cerebral blood flow pattern, with an increased perfusion in the mesial temporal lobe structures and decreased perfusion in the posterior cingulate gyrus. The discharges involved mainly the temporal lobe structures at seven sites (Fig. 3) and the temporoparieto-occipital junction at one site (Fig. 4).
|
|
|
Of the remaining four discharges, all of temporal origin, three did not exhibit any CBF increase and one resulted in a pronounced focal CBF decrease (Fig. 5
). The three discharges showing no associated CBF changes were restricted to the mesial temporal lobe. The discharge associated with a CBF decrease involved mainly the amygdala, where no CBF changes were observed; conversely, the hippocampus, where the CBF decreased (Fig. 5), was unaffected by the discharge, at least during the 45 s after discharge onset.
|
Because of the small amount of data, no correlation could be made between ictal CBF changes and other parameters such as the site and type of stimulation, the duration of discharge, the delay between tracer injection and discharge onset, the ictal symptoms and the presence of hippocampal atrophy (Tables 1 and 2
).
CBF–stereo-EEG anatomical correlates
In Fig. 6, the CBF findings per discharge are plotted versus stereo-EEG activity in the 10 anatomical regions likely to be involved by the spreading of temporal lobe discharges in our patients. We checked that no significant ictal CBF changes were observed elsewhere.
Fifty-nine of the 120 anatomical regions plotted in Fig. 6 were not explored by intracerebral electrodes, and 45 of the 59 (76%) demonstrated no ictal CBF changes. In 10 (17%), an ictal CBF increase (n = 9) or decrease (n = 1) was observed in the hippocampal gyrus (n = 6), the amygdala (n = 3) or the uncus (n = 1). These CBF changes were closely correlated with those observed concomitantly in the hippocampus, which was explored in all patients. In these cases, the spatial sampling of stereo-EEG, though not exhaustive, provided enough information to allow us to draw a conclusion as to whether the amygdalo-unco-hippocampal complex was affected by the discharge or not. The four remaining CBF changes observed in regions unexplored by stereo-EEG (7%) were an increase in CBF in the insula (n = 2) and temporal neocortex (n = 1) and a decrease in CBF in the posterior cingulate gyrus (n = 1). Although none of these structures was surgically removed, three patients were seizure-free after surgery while the remaining patient continued to experience seizures because of the incomplete removal of the epileptogenic zone (Table 1). Thus, these variations in CBF in regions unexplored by the electrodes did not prove relevant in predicting the surgical outcome of cortectomy tailored on the basis of stereo-EEG data.
Among the 61 stereo-EEG recorded structures, CBF measurements closely correlated with depth EEG findings in 43 (70.5%). We considered that CBF and stereo-EEG were correlated either when a significant increase (n = 19) or decrease (n = 1) in CBF was detected within a discharging area, or when a structure not affected by the induced discharge exhibited no CBF change (n = 23). A discrepancy between stereo-EEG and PET was found in 18 of the 61 sites explored by stereo-EEG (29.5%), among which stereo-EEG missed only one region of ictal CBF increase, located within the temporal neocortex. Thus, discrepancies between stereo-EEG and PET were mainly at discharging sites that showed no ictal CBF changes (17/18), namely, in various combinations, the temporopolar region (n = 5), the mesiotemporal lobe structures (n = 9) and the posterior cingulate gyrus (n = 3). This last structure was the only one that never demonstrated a CBF increase during electrically induced discharges.
When considering only the eight discharges accompanied by increased CBF, the percentage of concordance reached 82% (32 of the 39 stereo-EEG explored areas). It should be noted that all but one of the 20 sites of increased CBF explored by stereo-EEG electrodes (95%) proved to be electrically involved. However, PET failed to identify six of the 25 discharging sites (24%). Electrical changes undetected by PET were observed in the amygdala (n = 1), temporal pole (n = 2) and posterior cingulate gyrus (n = 3).
Influence of the ictal electrical pattern
Figure 7 sums up, case by case, the electrical patterns recorded at each site involved during the course of induced discharges, their evolution over time and the corresponding CBF changes.
|
The temporal resolution of PET being much lower than that of stereo-EEG, the question whether ictal CBF changes could be influenced by the type of electrical discharge was difficult to address. Furthermore, the 50-Hz stimulation produces a stimulus artefact that can impair the analysis of early ictal electrical activity.
However, considering each discharge as a whole, a CBF increase was observed during six of the nine discharges characterized at onset by a low-voltage fast activity or a high-voltage fast-spike train. Moreover, the four largest ictal CBF increases (>40%) were observed during discharges, beginning with a low-voltage fast activity, associated with high-frequency spiking. The greatest CBF increase (55%) that we observed was detected during the only discharge that exhibited exclusively these two electrical patterns, in spite of its particularly short duration (8 s). This suggests that a low-voltage fast activity could be one of the factors determining the extent of the increase in CBF, and that a rhythmic spiking activity is less relevant.
Nevertheless, a prolonged fast activity of low amplitude occurred in the absence of CBF changes during one discharge (case 8), and was associated with a major CBF decrease in another (case 9). The discharge remained restricted to the hippocampal and parahippocampal structures in the first of these two cases, and involved only the amygdala during the first 20 s from onset in the second case.
Two patients experienced two discharges of similar pattern (rhythmic spiking activity) and location; only one of these discharges was associated with a detectable increase in CBF. In these two patients, only the longer of the two seizures produced a detectable increase in CBF.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Most of the well-documented cases of ictal CBF changes, if not all of them, have revealed that CBF increases during the clinical seizure, closely following the discharge onset (Penfield et al., 1939
; Jasper and Erickson, 1941; Dymond and Crandall, 1976; Johnson et al., 1993; Weinand et al., 1994; Wallstedt et al., 1995; Spanaki-Varelas et al., 1997). Our H215O-PET data support these findings only partially, since three mesiotemporal discharges were not accompanied by any CBF alterations (25%) and one (8.5%) demonstrated a very significant focal decrease in the cerebral perfusion. However, the expected focal hyperperfusion was observed in eight out of 12 induced discharges (66.5%), thus confirming the possibility of detecting early ictal CBF increases using functional neuroimaging techniques. Before discussing these results, we will address some methodological issues concerning the possible bias linked to the use of electrical stimulation, and the difficulty in correlating data from two techniques with different spatial and temporal resolutions such as stereo-EEG and PET.
Methodological aspects
Electrically induced discharges
During intracranial depth EEG procedures, electrical stimulation proved to be a useful and reliable tool with which to reproduce patients' seizures (Penfield and Jasper, 1954; Wieser et al., 1979; Bernier et al., 1990; Kahane et al., 1993; Munari et al., 1993).
As shown by Penfield and colleagues (Penfield et al., 1939) for cortical surface stimulation, depth cortical stimulation per se did not elicit any CBF changes in our study.
The induced discharges concerned the temporal lobe structures in all cases but one, mainly because of the selection criteria. Indeed, subclinical seizures, or seizures accompanied only by subjective symptoms, are much more frequently encountered in temporal lobe epilepsy. In addition, electrical stimulation is more effective in triggering epileptic discharges in temporal lobe epilepsy (Munari et al., 1993) than in other types of partial epilepsy, probably because of the particularly low afterdischarge threshold of the amygdalohippocampal complex. As a matter of fact, stimulations that induced a focal discharge were applied within the hippocampus or the amygdala in 10 of 12 cases.
Spatial sampling errors
Using stereo-EEG, the sampling error concerns spatial resolution, since recordings are continuous in time but use a limited number of electrodes picking up the signal between paired leads that explore only a very restricted part of the cortex. In this study, 24% of sites showing ictal CBF changes were not explored by intracerebral electrodes. However, when compared with stereo-EEG, ictal PET data provided little additional information for the analysis of seizure spread. The main additional information provided by PET was that both the hippocampus and the amygdala showed ictal CBF increase when only one of these structures was explored by stereo-EEG and showed a focal discharge.
Temporal sampling errors
Reliable localization of the structures involved in an epileptic seizure requires the demonstration of a focal physiological change that takes place at the same time as the electroclinical event. In relation to this, PET using H215O explores the entire volume of the brain but has limited time resolution, since images are averaged over a period of 90 s. Little is known about the dynamic aspects of ictal cerebral perfusion changes, particularly at seizure onset. Weinand and colleagues (Weinand et al., 1994) demonstrated that CBF increased from baseline as long as 20 min before the occurrence of a seizure, abruptly declined at seizure onset, and thereafter quickly increased in a biphasic manner. Conversely, Wallstedt and colleagues (Wallstedt et al., 1995) showed that seizures could be preceded by a slow decrease in cortical perfusion that was maximal a few seconds before the seizure onset, and then showed the expected postictal hyperaemia. Thus, CBF PET images are likely to reflect a mixture of all these changes, depending on the timing of tracer injection after seizure onset. This could at least partially explain why CBF can remain unchanged, or even decrease, notably if the tracer is injected before the discharge onset.
Ictal CBF increase
The present H215O-PET study demonstrates that a marked CBF increase (16–55% above baseline) can be detected at the onset of ictal discharges arising from temporal or extratemporal areas, even in the absence of ictal symptoms (Table 2). This could explain some reports of `interictal' focal hyperperfusion using SPECT (Bluestone et al., 1989; Duncan et al., 1990), possibly due to subclinical seizures undetected on scalp EEG.
Globally, the magnitude of the CBF changes does not seem to correlate with the duration of discharges in this study, and the most prominent increase was found during the course of the shortest ictal event recorded, which lasted <10 s. These findings do not correlate with those of Johnson and colleagues (Johnson et al., 1993), who demonstrated a direct relationship between the duration of the afterdischarge and the extent of the focal CBF increase. A possible explanation is that, in our patients, the discharges began during the first seconds of uptake of the flow tracer, but other factors could be involved, notably the type of electrical pattern recorded at onset.
Areas of increased CBF, when explored by stereo-EEG, indicated an underlying epileptic discharge in 95% of the cases, so that CBF changes observed in unexplored areas were also likely to reflect an ictal discharge. Thus, PET provides information complementary to that provided by stereo-EEG, leading to better understanding of the `epileptogenic' network spatial distribution, and pathways of discharge propagation. Our data confirmed that ictal CBF changes were rarely limited to the hippocampus, but frequently involved both the amygdalohippocampal complex and the inner portion of the temporal pole (Brodmann area 38). They also suggest that some temporal lobe discharges, often mentioned but rarely demonstrated, can rapidly propagate to the insular cortex.
However, we observed neither the widespread ictal CBF increase frequently reported when using SPECT in so-called mesial temporal lobe seizures nor its usual association with hypoperfusion of the surrounding cortex and/or of the whole ipsilateral frontal lobe or hemisphere (Duncan, 1997b). A possible explanation is that the induced discharges were at their onset when the uptake of tracer began, and remained limited in space. Another possibility is that our conservative statistical criteria of significant CBF differences between conditions underestimated the extent of ictal CBF changes that could have been detectable visually on subtraction images of ictal minus rest conditions. However, our data analysis procedure is validated by the frequent co-localization observed between ictal stereo-EEG and PET CBF changes (76%) when considering the stereo-EEG areas explored that were involved in the induced discharges.
PET failed to identify almost a quarter of the discharging structures; of this quarter, half concerned the posterior cingulate gyrus. This relatively high frequency of falsely negative PET findings within a structure such as the posterior cingulate gyrus could be due either to difficulty in exploring the medial hemispheric cortex with PET or to the fact that the degree of ictal CBF change may vary among cortical areas.
The combination of hyper- and hypoperfusion was rare in our series, possibly because of the limited extent of ictal discharges, which consequently produced CBF increases that were too localized to create a blood volume shunt for discharging areas.
Absence of CBF changes during electrically induced focal epileptic discharges
One of the unexpected findings of our study was that CBF remained unchanged during three of the 12 electrically induced discharges (25%), all of which arose from the mesial temporal cortex. It is well known in clinical practice that SPECT functional imaging techniques for detecting CBF have lower sensitivity in extratemporal than in temporal seizures (Spencer, 1994); notably, some frontal lobe seizures may simply be too short to be detected by this method. However, in mesiotemporal lobe seizures, true ictal SPECT has been reported to show ictal CBF changes in almost all cases (Berkovic et al., 1993; Newton et al., 1995). Our data suggest that when the discharge originates in and remains restricted to mesiotemporal areas, with no involvement of the outer portion of the temporal pole (Brodmann area 20) or the temporal neocortex, it can remain undetected by CBF studies. The question remains as to whether such discharges are not detectable because of the resolution of PET images or because CBF changes are too small to reach statistical significance compared with the interictal condition.
CBF decrease during electrically induced focal epileptic discharges
Ictal CBF changes leading only to focal hypoperfusion have been reported, mainly in the course of frontal lobe seizures (Duncan, 1997b). A puzzling finding of our study was that a mesiotemporal lobe discharge resulted exclusively in a major decrease (64%) in cerebral perfusion. It has been shown that a fall of CBF can be observed far from the seizure site (Penfield et al., 1939), and that optical changes, which might reflect changes in blood volume, shifted below the baseline in areas surrounding an afterdischarge activity (Haglund et al., 1992). The CBF decrease can occur at seizure onset (Weinand et al., 1994) or immediately before a seizure (Wallstedt et al., 1995), with changes that greatly vary from sites only 10 mm apart (Wallstedt et al., 1995). In our patient showing a pure CBF ictal decrease, one may postulate that the area that was involved at discharge onset was too limited in space to produce a detectable CBF increase, but was surrounded by a larger area of inhibition producing a PET detectable CBF decrease.
|
Acknowledgments |
|---|
We wish to thank Dr D. Hoffmann for the placement of intracranial electrodes, Dr L. Minotti for the management of patients and Mrs E. Gamblin and V. Balle for their technical assistance. We are grateful to J. B. Poline for methodological help. This work was supported by grants from the Délégation Régionale à la Recherche Clinique du Centre Hospitalier Universitaire de Grenoble, the Région Rhône-Alpes, and the Programme Hospitalier de Recherche Clinique (PHRC 1994).
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionReferences |
|---|
Bancaud J, Talairach J, Bonis A, Schaub C, Szikla G, Morel P, et al. La stéréo-électroencéphalographie dans l'épilepsie. Informations neurophysiopathologiques apportées par l'investigation fonctionnelle stéréotaxique. Paris: Masson; 1965.
Benabid AL, Lavallée S, Hoffmann D, Cinquin P, Demongeot J, Danel F. Computer-driven robot for stereotaxic neurosurgery. In: Kelly PJ, Kall BA, editors. Computers in stereotaxic neurosurgery. Boston: Blackwell Scientific Publications; 1992. p. 330–42.
Berkovic SF, Newton MR, Chiron C, Dulac O. Single photon emission computed tomography. In: Engel J Jr, editor. Surgical treatment of the epilepsies. 2nd ed. New York: Raven Press; 1993. p. 233–43.
Bernier GP, Richer F, Giard N, Bouvier G, Mercier M, Turmel A, et al. Electrical stimulation of the human brain in epilepsy. Epilepsia 1990; 31: 513–20.[Web of Science][Medline]
Bluestone DL, Engelstad BL, Barbaro NM, Laxer KD. Technetium HMPAO-SPECT and high-resolution MRI in mesial temporal sclerosis [abstract]. Ann Neurol 1989; 26: 464.
Duncan JS. Imaging and epilepsy. [Review]. Brain 1997a; 120: 339–77.
Duncan R. The clinical use of SPECT in focal epilepsy. [Review]. Epilepsia 1997b; 38 Suppl 10: 39–41.
Duncan R, Patterson J, Hadley DM, Macpherson P, Brodie MJ, Bone I, et al. CT, MR and SPECT imaging in temporal lobe epilepsy [see comments]. J Neurol Neurosurg Psychiatry 1990; 53: 11–5. Comment in: J Neurol Neurosurg Psychiatry 1991; 54: 1030.
Duncan R, Patterson J, Roberts R, Hadley DM, Bone I. Ictal/postictal SPECT in the pre-surgical localisation of complex partial seizures. J Neurol Neurosurg Psychiatry 1993; 56: 141–8.
Dymond A, Crandall PH. Oxygen availability and blood flow in the temporal lobes during spontaneous epileptic seizures in man. Brain Res 1976; 102: 191–6.[Web of Science][Medline]
Engel J Jr. Outcome with respect to epileptic seizures. In: Engel J Jr, editor. Surgical treatment of the epilepsies. New York: Raven Press; 1987. p. 553–71.
Franck G, Sadzot B, Salmon E, Depresseux JC, Grisar T, Peters JM, et al. Regional cerebral blood flow and metabolic rates in human focal epilepsy and status epilepticus. Adv Neurol 1986; 44: 935–48.[Medline]
Gibbs FA. Cerebral blood flow preceding and accompanying experimental convulsions. Arch Neurol Psychiatry 1933; 30: 1003–10.
Gibbs FA, Lennox WG, Gibbs EL. Cerebral blood flow preceding and accompanying epileptic seizures in man. Arch Neurol Psychiatry 1934; 32: 257–72.
Grünwald F, Menzel C, Pavics L, Bauer J, Hufnagel A, Reichmann K, et al. Ictal and interictal brain SPECT imaging in epilepsy using technetium-99m-ECD. J Nucl Med 1994; 35: 1896–901.
Haglund MM, Ojemann GA, Hochman DW. Optical imaging of epileptiform and functional activity in human cerebral cortex. Nature 1992; 358: 668–71.[Medline]
Henry TR. Functional neuroimaging with positron emission tomography [Review]. Epilepsia 1996; 37: 1141–54.[Web of Science][Medline]
Herscovitch P, Markham J, Raichle ME. Brain blood flow measured with intravenous H215O. I. Theory and error analysis. J Nucl Med 1983; 24: 782–9.
Ho SS, Berkovic SF, McKay WJ, Kalnins RM, Bladin PF. Temporal lobe epilepsy subtypes: differential patterns of cerebral perfusion on ictal SPECT. Epilepsia 1996; 37: 788–95.[Web of Science][Medline]
Ingvar DH. rCBF in focal cortical epilepsy [abstract]. Stroke 1973; 4: 359.
Jasper H, Erickson TC. Cerebral blood flow and pH in excessive cortical discharge induced by metrazol and electrical stimulation. J Neurophysiol 1941; 4: 333–47.
Johnson DW, Hogg JP, Dasheiff R, Yonas H, Pentheny S, Jumao-as A. Xenon/CT cerebral blood flow studies during continuous depth electrode monitoring in epilepsy patients. AJNR Am J Neuroradiol 1993; 14: 245–52.[Abstract]
Kahane P, Tassi L, Francione S, Hoffmann D, Lo Russo G, Munari C. Manifestations électrocliniques induites par la stimulation électrique intracérébrale par `chocs' dans les épilepsies temporales. Neurophysiol Clin 1993; 22: 305–26.
Kuhl DE, Engel J Jr, 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: 348–60.[Web of Science][Medline]
Munari C, Bancaud J. The role of stereo-electro-encephalography (SEEG) in the evaluation of partial epileptic patients. In: Porter RJ, Morselli PL, editors. The epilepsies. London: Butterworths; 1987. p. 267–306.
Munari C, Kahane P, Tassi L, Francione S, Hoffmann D, Lo Russo G, et al. Intracerebral low frequency electrical stimulation: a new tool for the definition of the `epileptogenic area'? Acta Neurochir (Wien) 1993; Suppl 58: 181–5.
Munari C, Hoffmann D, Francione S, Kahane P, Tassi L, Lo Russo G, et al. Stereo-electroencephalography methodology: advantages and limits. Acta Neurol Scand 1994; Suppl 152: 56–67.
Newton MR, Berkovic SF, Austin MC, Rowe CC, McKay WJ, Bladin PF. SPECT in the localisation of extratemporal and temporal seizure foci. J Neurol Neurosurg Psychiatry 1995; 59: 26–30.
Penfield W. The evidence for cerebral vascular mechanism in epilepsy. Ann Intern Med 1933; 7: 303–10.
Penfield W, Jasper H, editors. Epilepsy and the functional anatomy of the human brain. London: J & A Churchill; 1954.
Penfield W, von Santha K, Cipriani A. Cerebral blood flow during induced epileptiform seizures in animals and man. J Neurophysiol 1939; 2: 257–67.
Poline JB, Mazoyer BM. Analysis of individual brain activation maps using hierarchical description and multi-scale detection. IEEE Trans Med Imaging 1994; 13: 702–10.[Web of Science][Medline]
Spanaki-Varelas M, Zubal G, MacMullan J, Spencer SS. Periictal SPECT localization verified by simultaneous intracranial EEG [abstract]. Epilepsia 1997; 38 Suppl 8: 62.
Spencer SS. The relative contributions of MRI, SPECT, and PET imaging in epilepsy. [Review]. Epilepsia 1994; 35 Suppl 6: S72–89.
Talairach J, Bancaud J. Stereotaxic approach to epilepsy. Methodology of anatomo-functional stereotaxic investigations. Prog Neurol Surg 1973; 5: 297–354.
Talairach J, Tournoux P, editors. Co-planar stereotaxic atlas of the human brain: 3-dimensional proportional system : an approach to cerebral imaging. Stuttgart: Thieme; 1988.
Talairach J, Bancaud J, Szikla G, Bonis A, Geier S, Vedrenne C. Approche nouvelle de la neurochirurgie de l'épilepsie. Méthodologie stéréotaxique et résultats thérapeutiques. Neurochirurgie 1974; 20: 1–240.
Theodore WH, Balish M, Leiderman D, Bromfield E, Sato S, Herscovitch P. Effect of seizures on cerebral blood flow measured with 15O-H2O and positron emission tomography. Epilepsia 1996; 37: 796–802.[Web of Science][Medline]
Trébossen R, Mazoyer BM. Some physical characteristics of a time-of-flight positron tomography (CEA-LETI-TTVO3) obtained with the EEC emission phantom. Med Prog Technol 1991; 17: 165–71.[Web of Science][Medline]
Wallstedt L, Gazelius B, Lind G, Meyerson BA, Linderoth B. Chronic multifocal recording of cortical microcirculation and subdural EEG during epileptic seizures in humans [abstract]. Epilepsia 1995; 36 Suppl 3: S146–7.
Weinand ME, Carter LP, Patton DD, Oommen KJ, Labiner DM, Talwar D. Long-term surface cortical cerebral blood flow monitoring in temporal lobe epilepsy. Neurosurgery 1994; 35: 657–64.[Web of Science][Medline]
Wieser HG, Bancaud J, Talairach J, Bonis A, Szikla G. Comparative value of spontaneous and chemically and electrically induced seizures in establishing the lateralization of temporal lobe seizures. Epilepsia 1979; 20: 47–59.[Web of Science][Medline]
Woods RP, Cherry SR, Mazziotta JC. Rapid automated algorithm for aligning and reslicing PET images. J Comput Assist Tomogr 1992; 116: 620–33.
Woods RP, Mazziotta JC, Cherry SR. MRI–PET registration with automated algorithm. J Comput Assist Tomogr 1993; 17: 536–46.[Web of Science][Medline]
Received January 8, 1999. Revised April 14, 1999. Accepted April 22, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. Rubboli, R. Mai, S. Meletti, S. Francione, F. Cardinale, L. Tassi, G. L. Russo, M. Stanzani-Maserati, G. Cantalupo, and C. A. Tassinari Negative myoclonus induced by cortical electrical stimulation in epileptic patients Brain, January 1, 2006; 129(1): 65 - 81. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Zumsteg, R. A. Wennberg, V. Treyer, A. Buck, and H. G. Wieser H215O or 13NH3 PET and electromagnetic tomography (LORETA) during partial status epilepticus Neurology, November 22, 2005; 65(10): 1657 - 1660. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Chabardes, P. Kahane, L. Minotti, L. Tassi, S. Grand, D. Hoffmann, and A. L. Benabid The temporopolar cortex plays a pivotal role in temporal lobe seizures Brain, August 1, 2005; 128(8): 1818 - 1831. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kaminska, C. Chiron, D. Ville, G. Dellatolas, A. Hollo, C. Cieuta, C. Jalin, O. Delalande, M. Fohlen, P. Vera, et al. Ictal SPECT in children with epilepsy: comparison with intracranial EEG and relation to postsurgical outcome Brain, January 1, 2003; 126(1): 248 - 260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Kovacs, S. Schuchmann, S. Gabriel, O. Kann, J. Kardos, and U. Heinemann Free Radical-Mediated Cell Damage After Experimental Status Epilepticus in Hippocampal Slice Cultures J Neurophysiol, December 1, 2002; 88(6): 2909 - 2918. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Bouilleret, M. P. Valenti, E. Hirsch, F. Semah, and I. J. Namer Correlation Between PET and SISCOM in Temporal Lobe Epilepsy J. Nucl. Med., August 1, 2002; 43(8): 991 - 998. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Ostrowsky, M. Magnin, P. Ryvlin, J. Isnard, M. Guenot, and F. Mauguiere Representation of Pain and Somatic Sensation in the Human Insula: a Study of Responses to Direct Electrical Cortical Stimulation Cereb Cortex, April 1, 2002; 12(4): 376 - 385. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Van Paesschen, P. Dupont, B. Van Heerden, H. Vanbilloen, L. Mesotten, A. Maes, G. Van Driel, and L. Mortelmans Self-injection ictal SPECT during partial seizures Neurology, May 23, 2000; 54(10): 1994 - 1997. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||