EEG Background Delta Activity in Temporal Lobe Epilepsy: Correlation with Volumetric and Spectroscopic Imaging

Transcription

1 Epilepsia, 40(11):158&1586, 1999 Lippincott Williams & Wilkins, Inc., Philadelphia 0 International League Against Epilepsy Clinical Research EEG Background Delta Activity in Temporal Lobe Epilepsy: Correlation with Volumetric and Spectroscopic Imaging Andrea Bernasconi, Fernando Cendes, Jong Lee, D. C. Reutens, and Jean Gotman Department of Neurology, Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec, Canada Summary: Purpose: With quantitative electroencephalogram (EEG) and neuroimaging methods, we examined delta activity, atrophy, and neuronal-axonal dysfunction of the cerebral gray and white matter in patients with intractable temporal lobe epilepsy (TLE). Based on evidence that lesions of the white matter result in EEG delta activity, we postulated that background abnormalities in patients with TLE are related to changes of the temporal lobe white matter. Methods: We measured interictal delta activity in 34 TLE patients and 10 controls. Spike-free and artifact-free EEG samples were selected by visual inspection. A spectral analysis was used to compute the energy in the delta frequency band. We compared the results of the spectral analysis to magnetic resonance imaging- (MRI) based volumes of the temporal lobe white and gray matter, the hippocampus and the amygdala; and N-acetyl aspartate (NAA) in the lateral and posterior temporal lobe by using proton magnetic resonance spectroscopic imaging ('H-MRSI). The degree of correlation between delta activity and the neuroimaging measurements was assessed by using the Pearson correlation coefficient (r). An analysis of variance (ANOVA) was used to examine the influence of the seizurefocus lateralization on the delta activity and the neuroimaging parameters. Results: There was no significant difference in the amount of delta activity in the temporal lobe between the controls and patients. We found no correlation between delta activity and the neuroimaging measures (p > 0.05). The ANOVA showed significant differences between the patients and controls for the volume of the gray and white matter of the temporal lobe and for the NAA in the lateral and posterior temporal lobe (p < 0.002). Conclusions: The interictal background delta activity was not explained by reduced volume of the temporal lobe white matter, gray matter, or by abnormalities seen in 'H-MRSI. Key Words: Delta waves-eeg spectral analysis-proton magnetic resonance spectroscopy-magnetic resonance imaging- Temporal lobe epilepsy. Delta activity (0-4 Hz) can be intermittenr or intermingled with the electroencephalogram (EEG) background and is more or less continuous. The mechanisms underlying the generation of pathologic delta activity are not clear. However, clinical and experimental observations indicate that continuous polymorph delta activity results primarily from lesions affecting cerebral white matter. It has been proposed that localized lesions of the white matter in cats produce focal polymorph delta waves resulting from partial cortical deafferentation from subcortical inputs, with the delta waves generated by relatively normal cortical neurons (1). With microphysiologic techniques, Ball et al. (2) postulated that the generator of delta waves is most likely located in cortical Accepted April 8, Address correspondence and reprint requests to Dr. A. Bemasconi at Montreal Neurological Institute, 3801 University Street, Montreal, PQ, Canada. pyramidal neurons. EEG patterns dominated by irregular delta waves can be found in the athalamic cat (3) and in the isolated cerebral hemisphere (4). In seizure disorders, interictal focal slow waves often suggest local lesions such as tumors, strokes, injuries, or neuronal migration disorders. Focal slow waves also can be seen in nonlesional epilepsies as part of ongoing epileptiform activity. However, the origin of EEG background abnormalities in patients with nonlesional intractable temporal lobe epilepsy (TLE) remains unclear. Computerized spectral analysis allows a quantification of abnormal delta activity in EEG background and is particularly effective when the EEG seems normal on visual inspection (5). Magnetic resonance-based volumetric imaging (MRIV) and proton magnetic resonance spectroscopic imaging ( 'H-MRSI) allow, respectively the quantitative and functional assessment of the epileptogenic cortex. In TLE, MRIV studies have demonstrated hippocampal and amygdalar atrophy ipsilateral to 1580

2 DELTA ACTIVITY IN TEMPORAL LOBE EPILEPSY 1581 the seizure focus (6). By using a three-dimensional volumetric analysis, we (7) and others (8) also demonstrated a bilateral decrease in the volume of gray matter and a reduction of the white matter volume ipsilateral to the seizures focus. 'H-MRSI allows for the determination of a reduced signal intensity of N-acetyl aspartate (NAA) and a creatine (Cr) resonance intensity from multiple cerebral regions simultaneously. The metabolites Cr and NAA are both found within the cortex and the white matter of human brain, with NAA found in higher concentrations in the gray matter than in the white (9). 'H- MRS revealed a relatively low NAA/Cr resonance intensity in the middle and lateral temporal lobes of patients with TLE ipsilateral to the EEG focus, which implies a loss or dysfunction of neurons and their processes (10,ll). The purpose of our study was to test the hypothesis that EEG background abnormalities in TLE are related to changes of the white matter of the temporal lobe and, alternatively, to test whether EEG background abnormalities in TLE are related to damage in the gray matter. METHODS Subjects We studied 34 patients with medically intractable TLE (10 men and 24 women; mean age, 36 years). The type and site of seizure onset were determined by a comprehensive evaluation including a detailed history, neurologic examination, review of medical and EEG records, and neuropsychologic evaluation. Data on age at seizure onset, duration of the current seizure disorder, and estimated monthly seizure frequency were obtained from direct interviews with the patients and from medical charts. Duration of the seizure disorder was calculated as the difference between the age at which the MRI scan was performed and the age at seizure onset. The mean age of the first afebrile seizure was 12 years; the mean duration of epilepsy was 24 years; and the mean seizure frequency was 18 seizures per month. Thirteen patients had a history of febrile convulsions in infancy. Patients with focal mass lesions were not included. The epileptic focus was determined by predominant interictal epileptic abnormalities, unequivocal seizure onset recorded during prolonged video-eeg monitoring by using sphenoidal electrodes, and, in 26 cases, response to surgical treatment. Patients also were evaluated with neuropsychological testing and intracarotid amobarbital testing for language and memory localization. Seventeen patients underwent a selective amygdalohippocampectomy, and nine underwent a temporal cortical resection with amygdalohippocampectomy. The postoperative qualitative histopathology showed hippocampal sclerosis (12) in all patients. Based on these criteria, patients with TLE were divided into those with a left-sided (n = 13) or right sided (n = 21) seizure focus. EEG recordings and spectral analysis EEG recordings were performed for all 34 patients and (the 10 control subjects (two men and eight women; mean age, 32 years). Interictal background activity was studied and selected by visual inspection in 4-min EEG samples recorded the first day of the telemetry period. A spike-free and artifact-free EEG sample of 1 min was considered to be a satisfactory sample of the background activity (long enough to be representative, yet short enough to remain stationary). Recordings were performed with scalp electrodes placed according to the International system with sphenoidal or zygomatic electrodes (on the zygomatic arch) and supplementary electrodes in F,/Flo, P,/P,,, and T,/TIo. Samples were obtained while the subject was awake and relaxed, with eyes closed. The digitizing rate was set at 200 samples/s/channel. A biocalibration was recorded before sampling. To avoid postictal spike activation or slow waves, only samples obtained 212 h after an ictal phenomenon (complex partial seizure or secondarily generalized tonic-clonic seizures) were considered. Spectral analysis was performed on a bipolar montage by computing the Fast Fourier Transform on epochs of 2.56 seconds, which were then averaged over the selected sections. Background abnormalities were evaluated by using the energy in the delta band (1-3.9 Hz). Results were compared with 10 neurologically normal control subjects diagnosed with pseudoseizures (nonepileptic seizures) (13). All had several, previously normal, standard scalp EEG recordings. Episodes of loss of responsiveness recorded in combination with normal EEG readings during the video-eeg telemetry recordings were representative of those that occurred spontaneously in the control subjects at home. A normal high-resolution MRI ruled out diffuse encephalopathy or cortical lesions. Family history of epilepsy was negative in all control subjects. EEG samples for quantification of delta activity were carried out according to the protocol described for the TLE patients. The mean value of the absolute (in pv) and relative activity of the delta band were calculated in the following channels: F,/F,-T,/T,, T,/T,-TS/T6, Ts/T6-0 '/02, F,/F, o-t,/t, o, T,/T, "-P,/P, o, P9/P,,-0, /02. To avoid artifacts caused by eyeball movements, delta activity was not calculated in channels F,,-F9 and FP2-Flo. The parasagittal chains (F /F2, F3/F4, C,/C,, P,/P,, 0,/02 elecp.' trodes) were not included because they do not reflect activity in the temporal lobes. MRIV of the hippocampus and amygdala All patients also underwent hippocampal and amygdalar volumetry by using a previously described method (6). Volumes were corrected for the variation of intracranial volume and compared with normal control data to allow detection of bilateral atrophy. Compared with the Epilepsiu, Vol. 40, No. 11, I999

3 1582 A. BERNASCONI ET AL. controls, the mean volume of the hippocampus was significantly smaller ipsilateral to the seizure focus in the patients with TLE (p < 0.05). MRIV of the temporal neocortex and white matter Volumetry of the gray and white matter of the temporal lobe was carried out in 29 patients and 28 control subjects (17 men and nine women; mean age, 27 years) according to a previously published protocol (7). In brief, images were acquired on a 1.5 T gyroscan (Philips Medical System, Eindhoven, The Netherlands), by using a T,-fast field echo (repetition time [TR] 18, echo time [TE] 10; 1 acquisition average pulse sequence; flip angle, 30"; matrix size, 256 x 256; voxel size, 1 mm x 1 mm x 1 mm). Images were registered into stereotaxic space, corrected for image intensity and inhomogeneity, and automatically segmented into gray and white matter and cerebrospinal fluid over a predetermined extent of the temporal lobe. The boundaries of the volume of interest within each temporal lobe were demarcated on the coronal MR images using an interactive mouse-driven software. 'H-MRSI 'H-MRSI was performed in 33 patients and 21 neurologically normal controls (13 men and eight women, mean age, 28 years) according to a previous published protocol (1 0). MRI scans were acquired by using a 1.5-T combined imaging and spectroscopy system (Philips Medical Systems, Best, The Netherlands). Multislice spin-echo MRIs (TR, 2,000 ms; TE, 30 ins) were obtained in the transverse plane along the axis of the temporal lobes and the coronal plane perpendicular to the axis of the sylvian fissure. A large region of interest (ROI) behind the clivus that included both temporal lobes was defined for selective excitation before phase encoding for the 'H-MRSI. The ROI was oriented in a similar position for all examinations to cover the entire extension of both hippocampi. The values for NAA, choline (Cho), Cr, and NAA/Cr were determined for the middle, posterior, and lateral regions of each temporal lobe by averaging values from spectra (12 3) in these regions. (Fig. 1.) The middle temporal ROI comprises the head and body of the hippocampus. The posterior temporal ROI includes white matter of the temporal lobe and the tail of the hippocampus. Because of technical limitations, including lipid signal contamination, the lateral temporal region used for MRSI analysis comprised a limited portion of the lateral temporal lobe and included signals from both the gray and the white matter. Resonance intensities on the MRSIs were determined from peak areas by integration by using locally developed software. Spectra were excluded from the analyses if they were broadened by artifacts (i.e., full width with at half-maximum >10 Hz) or if the Cho or Cr peaks were FIG. 1. Proton magnetic resonance spectroscopic imaging (MRSI) region of interest (ROI). Conventional (water-based) axial T,-weighted MRI through both temporal lobes from a normal control subject with the original phase-encoding grid superimposed. The thick white line outlines the ROI from the acquisition of the MRSI. Each temporal lobe inside the ROI is subdivided into three anatomic regions. The most anterior (dotted black line), defined as middle temporal, comprises the region of the head and body of the hippocampus. The second, defined as posterior temporal (black line), includes the tail of the hippocampus and the white matter of the posterior temporal lobe. The lateral temporal region (dotted white line) comprises a limited portion of the lateral temporal lobe, and includes the signal from both the gray and white matter. To avoid artifacts from the bone and fat tissue from the clivus and sphenoid bone, the pole of the temporal lobe and the amygdala are not included in the ROI. not resolved. The spectroscopist was unaware of the results from the EEG investigation. The intensity NAA/Cr ratio was used to simplify quantitation across patients. The observations derived from the NAA/Cr ratio do not in themselves depend on Cr being a stable internal reference; however, if Cr is stable or only slightly increased in the brain regions associated with epileptogenic damage (14,15), observed decreases in the NAA/Cr ratio can be interpreted in terms of neuronal or axonal damage (10). Statistical analysis Because the values of delta activity did not exhibit a normal gaussian distribution, they were transformed into logarithmic values (16). Z scores were obtained for each individual patient by subtracting the values of each parameter from the mean value obtained from the control group and dividing the result by the control group's standard deviation. To determine the relation between delta activity and the duration of epilepsy, seizure frequency, volume of the temporal lobe gray and white matter, and NAA/Cr in the lateral and posterior temporal lobe, we performed Pearson correlation analyses and calculated a correlation coefficient (r). An one-factor ANOVA was Epiltpsia, Vol. 40, No. 11, 1999

4 DELTA ACTIVITY IN TEMPORAL LOBE EPILEPSY 1583 used to examine the influence of the epileptic-focus lateralization on delta activity, gray and white matter volumes, hippocampal and amygdalar volumes, and NAA/ Cr values of the lateral and posterior temporal lobe in the patients with TLE and the controls. RESULTS A single EEG section was analyzed in seven control subjects and 19 patients. Because of ocular artifacts extending in the posterior channels in three control subjects and 15 patients, more than one section (mean, 4.7 for controls and 3.2 for patients) was analyzed. Mean duration of the sections was 12.9 s for the control subjects and 18.7 s for the TLE patients. Mean absolute values of delta activity in FV (& standard deviations) in the controls and patients are shown in Table 1. The mean delta activity was greater in the patients with right and left TLE than in the controls, but the difference was not statistically significant (F = 1.3 and p = 0.27). The mean volumes (& standard deviations) of the gray matter and white matter of the temporal lobe, the hippocampus, and the amygdala are shown in Table 2. ANOVA values for the white and gray matter volumes showed a significant difference between the control subjects and the patients with epilepsy (white matter, p = 0.002; gray matter, p < 0,001). ANOVA values for the hippocampus and amygdala volumes were significantly different in the TLE patients compared with the controls (hippocampus, p < 0.001; amygdala, p < 0.01). The NAA/Cr lateral and NAA/Cr posterior values (+ standard deviations) from both temporal lobes are shown in Table 3. The ANOVA for NAA/Cr in the lateral temporal lobe and NAA/Cr in the posterior temporal lobe showed a significant difference between the epilepsy patients and controls (p < 0.001). The mean values for the gray matter volume and NAA/Cr in the lateral temporal lobe were significantly different in the controls (p < 0.05, Table 2) for the right and left hemisphere. In each patient, these neuroimaging parameters were normalized by the mean of the controls for each side. We found no correlation between the average values of absolute delta activity with the volumes of the gray matter (r = 0.24; p = 0.07) and white matter TABLE 1. Mean f standard deviation delta activity in patients with TLE and controls Delta activity (pv) Patient group Right temporal lobe Left temporal lobe Right-side TLE patients Left-side TLE patients 9.2? Control5 4.8 & & 2.5 TLE, temporal lobe epilepsy. (Y = -0.05; p = 0.7) of the temporal lobe; the volume of the hippocampus (r = 0.12; p = 0.4) and the amygdala (r = -0.06; p = 0.6); and the NAA/Cr values in lateral (Y = 0.20; p = 0.2) and posterior (r = 0.30; p = 0.09) temporal regions (Fig. 2). No significant correlation could be shown between EEG delta activity and the duration of epilepsy or seizure frequency. DISCUSSION By using a multimodal neuroimaging approach, we found no correlation between the amount of delta activity and the structural or functional abnormalities of the temporal lobe, as expressed by the reduction in the volume of gray and white matter and the NAA/Cr ratio. We were unable to demonstrate the cause of the continuous background abnormalities in the delta frequency recorded within the temporal lobe of TLE patients. In our study, TLE patients were not selected based on the amount of delta activity. Even though the mean delta activity was greater in the patients than the controls, we found no statistically significant difference between the mean delta activity of patients compared with controls. This may be due to the large variability in the amount of delta activity in the TLE patients. We could have separately analyzed the patients who showed increased delta activity, but we wanted to retain a large range of delta activity, which would correspond with the wide range in neuroiinaging values to better reveal possible correlations. Correlation implies correspondence for high and low delta values. The biologic reason for the variability of EEG background abnormality in TLE patients remains unclear. One possible explanation for the low degree of delta activity found in some patients could be bilateral temporal gray matter damage, as demonstrated by the neuroimaging data (7) resulting in an amplitude reduction of the EEG signal. Damage to the gray matter decreases synchronization of active generators, which results in attenuated surface-recorded EEG amplitude (1,17). Injury to the white matter reduces the excitatory inputs to the neocortex, resulting in increased delta activity, which is generated in an intact cerebral cortex (1,2,18). In a study with cats, lesions of the white matter were circumscribed and carried out beneath the cortical surface, where myelinated axons are abundant. Nonparoxysmal delta waves were also predominantly observed in patients with diffuse encephalopathies with widespread and severe damage of the white matter (19). Therefore it is conceivable that even if the white matter abnormality in some of our patients was diffuse, it was not sufficiently severe to generate a large amount of delta waves. Thus our results do not exclude a priori an underlying minor and functional deafferentation of the cortex related to a white matter dysfunction, which did not manifest Epilepsiu, Vol. 40, No. 11, 19YY

5 1584 A. BERNASCONI ET AL. TABLE 2. Mean 2 standard deviation volume measurements of the gray and white matter of the temporal lobe, the hippocampus and the amygdala in patients with TLE and controls Right temporal lobe Volume measurements Left temporal lobe Hippocampus Amygdala Hippocampus Amygdala Patient group GM (cm ) WM (cm ) (mm ) (rnm ) GM (cm ) WM (cm ) (mm? (mm3) Right-side TLE patients & 3.1 3, & f 1.9 3, , Left-side TLE patients 34. I f f 1.8 3,993 f & k 2.4 2, ,103 i 280 Controls 40.0k f2.2 4, el , , TLE, temporal lobe epilepsy; GM, gray matter; WM, white matter. itself by measurable atrophy of the white matter or by a decrease of the NAA. It may be possible to improve the sensitivity of IH-MRSI by increasing spatial resolution by using a higher field-strength scanner (20), by developing multislice and fat-suppression H-MRSI techniques (21), or by using MRI-based tissue segmentation (22). The lack of correlation between the amount of delta activity and disease duration support the hypothesis that continuous EEG background abnormalities in TLE patients do not reflect a regional functional impairment secondary to chronic epileptogenesis. Prolonged AED therapy or short-term discontinuation withdrawal activities do not produce focal changes in EEG background delta activity (23-25). A neurotoxic action of the AED responsible for the generation of delta activity seems therefore improbable in our subjects. Could slow waves result from atrophy of deep mesial structures? Ipsilateral white matter volume reduction may reflect a loss of hippocampal projections to the temporal neocortex secondary to hippocampal neuronal loss. Given the positive correlation between the volume of the hippocampus and the white matter of the temporal lobe (r = 0.33; p < 0.02) in our patients, we were not surprised to find no correlation between continuous background delta activity and hippocampal or amygdalar atrophy. In a previous qualitative assessment (26), intracerebral EEG delta activity within the hippocampus and the amygdala was found to correlate significantly with atrophy of mesial temporal lobe structures. Obviously, slow waves recorded on the scalp do not reflect slow activity generated within the hippocampus or the amygdala. The difference in the two studies results is therefore understandable. A qualitative scalp EEG and neuroimaging study (27) also showed no correlation between interictal regional slow-wave activity and hippocampal atrophy as determined by visual MRI analysis. Another possible explanation is that slow waves originate from cortical deafferentation from the thalamus. Electophysiologic data from human and animals show that there is marked ipsilateral cortical EEG slowing after unilateral lesions or dysfunction of the nonsensory thalamic nuclei (I). Data from positron emission tomography (PET) studies (28) demonstrated hypometabolism in the basal ganglia and thalamus ipsilateral to the cortical seizure focus in patients with partial epilepsy, resulting in a focal disturbance of the subcortical regulatory function. Therefore the generation of slow waves in TLE patients could be related to a chronic functional deactivation of the lateral temporal neocortex due to a disruption in thalamocortical and corticolimbic projections (1,18,29). In the absence of cerebral anoxia, slow waves are thought to reflect a diminution in the cerebral oxidative metabolism usually associated with a diminution in cerebral blood flow (30). It is tempting to speculate that a dysfunction of vascular regulation (3 1) may contribute to chronic white matter hypoperfusion and, therefore, to the generation of delta activity. The small amount of delta activity in our TLE patients could parallel microvascular changes in the absence of a macroscopic visible lesion. In our group of patients, background delta activity was TABLE 3. Mean 2 standard deviation of N MCr values of the lateral and posterior temporal lobe in patients with TLE and controls Right temporal lobe NANCr values Left temporal lobe Patient group Lateral Posterior Lateral Posterior Right-side TLE patients 4.8 k & f 0.41 Left-side TLE patients 4.17 k f & 0.39 Controls f NANCr, N-acetyl aspartatekreatine ratio; TLE, temporal lobe epilepsy. Epilqnia, Vol. 40, No. 11, I999

6 DELTA ACTIVITY IN TEMPORAL LOBE EPILEPSY 1585 H : 3tr,;?& 5 L 0.60 v) t i: A delta (log) idelta (log) C G b e c E m &C c i, r ? - v 0.60 Ln a: I 0.55 R #% i; I I I I o o B delta (log) delta (log) D.,-. FIG -. 3 Srattnmram.,.,U'.",Y.U.,. hotwnnn ddit2.v.&u --"."I artiwitw -.,- and Y.UJ nraw...ul.i. matter \mil.","...",me (A). \-, white..,,,." mattnr \in11 imn IR\. mrnnetir rocnnanre snwtrnsrnnir lll-.lyl.-...,..-,-,,...- y.'" r----'----t'- imaging (MRSI) of the lateral temporal lobe (C) and the posterior temporal lobe (D) in patients with intractable temporal lobe epilepsy. Values were transformed into logarithm. t not explained by reduced volume of white or gray matter or by a reduction of NAA/Cr in the temporal lobe. The negative results could be related to a limitation of the MR techniques we used in detecting subtle structural or functional abnormalities of the cerebral white matter and do not exclude a subcortical functional deficit, which would lead to cortical deafferentation, probably of multifactorial origin. On the other hand, the pathophysiologic mechanism underlying the variability of the amount of delta waves in TLE patients remains to be explored. Thus EEG spectral analysis of delta activity may provide additional information independent of neuroimaging and pathologic data in patients with TLE. Acknowledgment: This work was supported in part by a grant (MT-10189) of the Medical Research Council of Canada and by the Savoy Foundation for Epilepsy Research, Montreal, Quebec, Canada. We thank Mme Allard and Mme Drouin for expert technical assistance. REFERENCES 1. Gloor P, Ball GJ, Schaul N. Brain lesions that produce delta waves in the EEG. Neurology 1977;27: Ball GJ, Gloor P, Schaul N. The cortical electromicrophysiology of pathological delta waves in the electroencephalogram of cats. Electroencephalogr Clin Neurophysiol 1977;43: Villablanca J. Role of the thalamus in sleep control: sleepwakefulness studies in chronic diencephalic and athalamic cats. In: Petre-Quadens 0, Schlag J, eds. Basic sleep mechanisms. New York: Academic Press, 1974: Kellaway P, Go1 A, Proler M. Electrical activity in the isolated cerebral hemisphere and isolated thalamus. Exp Neurol 1966; 14: Gotman J, Skuce DR, Thompson CJ, Gloor P, Ives JR, Ray WF. Clinical applications of spectral analysis and extraction of features from electroencephalograms with slow waves in adult patients. Electroencephalogr Clin Neurophysiol 1973;35: Cendes F, Andermann F, Gloor P, et al. MRI volumetric measurement of amygdala and hippocampus in temporal lobe epilepsy. Neurology 1993;43: Lee JW, Andermann F, Dubeau F, et al. Morphometric analysis of the temporal lobe in temporal lobe epilepsy. Epilepria 1998;39: Epilepsiu, Vol. 40, No. II, 1999

7 I586 A. BERNASCONI ET AL Marsh L, Morrel MJ, Shear PK, et al. Cortical and hippocampal volume deficits in temporal lobe epilepsy. Epilepsia 1997;38: Petroff OA, Spencer DD, Alger JR, Prichard JW. High-field proton magnetic resonance spectroscopy of human cerebrum obtained during surgery for epilepsy. Neurology 1989;39: Cendes F, Caramanos Z, Andermann F, Dubeau F, Arnold DL. Proton magnetic resonance spectroscopic imaging and magnetic resonance imaging volumetry in the lateralization of temporal lobe epilepsy: a series of 100 patients. Ann Neurol 1997;42: Hugg JW, Laxer KD, Matson GB, Maudsley AA. Weiner MW. Neuron loss localizes human temporal lobe epilepsy by in vivo proton magnetic resonance spectroscopic imaging. Ann Neurol 1993;34: Meencke HJ, Veith G. Hippocampal sclerosis in epilepsy. In: Liiders H, ed. Epilepsy surgery. New York: Raven Press, 1991: Lesser RP. Psychogenic seizures. Neurology I996;46: Connelly A, Jackson GD, Duncan JS, King D, Gadian DG. Magnetic resonance spectroscopy in temporal lobe epilepsy. Neurology 1994;44: Petroff OA, Pleban LA, Spencer DD. Symbiosis between in vivo and in vitro NMR spectroscopy: the creatine, N-acetylaspartate, glutamate, and GABA content of the epileptic human brain. Magn Reson Imaging 1995; 13: Gasser T, Baecher P, Moecks J. Transformations towards the normal distribution of broad band spectral parameters of the EEG. Electrorncephalogr Clin Neurophysiol 1982;53: Lopes da Silva FH. Neural mechanisms underlying brain-waves: from neural membranes to networks. Electroencephalogr Clin Neurophysiol 1991;79: Nakamura Y, Ohye C, Mano N. Cortical polarization and experimentally produced delta waves in the cat. Electroencephalogr Clin Neurophysiol 1968;24: Gloor P, Kalabay 0, Giard M. The electroencephalogram in diffuse encephalopathies: electroencephalographic correlates of grey and white matter lesions. Brain 1968;91: Hetherington H, Kuzniecky RI, Pan J, et al. Proton nuclear mag- netic resonance spectroscopic imaging of human temporal lobe epilepsy at 4.1 T. Ann Neurol 1995;38: Haupt C, Schuff N, Weiner M, Maudsley AA. Removal of lipid artifacts in 'H spectroscopic imaging by extrapolation. Magn Reson Med 1996;35: Michaelis T, Merboldt KD, Bruhn H, Hanicke W, Frahm J. Absolute concentrations of metabolites in the adult human brain in vivo: quantification of localised proton MR spectra. Radiology 1993; 187: van Wieringen A, Binnie CD, De Boer PTE, Van Emde Boas W, Overweg J, de Vries J. Electroencephalographic findings in antiepileptic drug trials: a review and report of 6 studies. Epilepsy Res 1987; Duncan JS, Smith SJ, Forster A, Shorvon SD, Trimble MR. Effects of the removal of phenytoin, carbamazepine and valproate on the electroencephalogram. Epilepsia 1989;30:59& Herkes GK, Lagerlund TD, Sharbrough FW, Eadie MJ. Effects of antiepileptic drug treatment on the background frequency of EEGs in epileptic patients. J Clin Neurophysiol 1993;10: Cendes F, Dubeau F, Andermann F, et al. Significance of mesial temporal atrophy in relation to intracranial ictal and interictal stereo EEG abnormalities. Brain : Koutroumanidis M, Binnie CD, Elwes RD, et al. Interictal regional slow activity in temporal lobe epilepsy correlates with lateral temporal hypometabolism as imaged with18fdg PET: neurophysiological and metabolic implications. J Neurol Neurosurg Psychiatry 1998;65: Sperling MR, Gur RC, Alavi A, et al. Subcortical metabolic alterations in partial epilepsy. Epilepsia 1990;31: Schaul N, Gloor P, Gotman J. The EEG in deep midline lesions. Neurology 1981;31: Ingvar DH, Sjolund B, Ardo A. Correlation between dominant EEG frequency, cerebral oxygen uptake and blood flow. Electroencephalogr Clin Neurophysiol 1976;41: Stan MS. The role of dopamine in epilepsy. Synapse 1996;22: Epilrpsia, Vol. 40, No. 11, 1999