- Email: [email protected]
Lack of aura experience correlates with bitemporal dysfunction in mesial temporal lobe epilepsy
Epilepsy Research 43 (2001) 201– 210 www.elsevier.com/locate/epilepsyres
Lack of aura experience correlates with bitemporal dysfunction in mesial temporal lobe epilepsy Reinhard Schulz a,*, Hans O. Lu¨ders b, Matthias Hoppe a, Hennric Jokeit a, Anke Moch a, Ingrid Tuxhorn a, Theodor May c, Alois Ebner a a
Abteilung fu¨r pra¨chirurgische Epilepsiediagnostik, Epilepsiezentrum Mara gGmbH, Epilepsiezentrum Bethel, Maraweg 21, 33617 Bielefeld, Germany b Cle6eland Clinic Foundation, Cle6eland, OH, USA c Gesellschaft fu¨r Epilepsieforschung, 33617 Bielefeld, Germany Received 27 July 2000; received in revised form 5 October 2000; accepted 15 October 2000
Abstract The diagnostic value of lack of aura experience in patients with temporal lobe epilepsy (TLE) is unclear. Purpose: To evaluate possible factors of bitemporal dysfunction in patients with mesial TLE who did not experience an aura in electroencephalography EEG/video monitoring for epilepsy surgery. Methods: Ictal scalp EEG propagation patterns of 347 seizures of 58 patients with mesial temporal lobe sclerosis or non-lesional mesial TLE, interictal epileptiform discharges (IED), presence of unilateral mesial temporal lobe sclerosis in visual magnetic resonance imaging (MRI) analysis, prose memory performance, history or not of an aura, and postictal memory or absence of an aura were analyzed. The ictal EEG was categorized as follows. EEG seizure: (a) remaining regionalized, (b) non-lateralized, (c) showing later switch of lateralization or bitemporal asynchronous ictal patterns. Results: Absent aura in monitoring was significantly correlated with absence of unitemporal MRI sclerosis (P= 0.004), bitemporal IED (P = 0.008), and propagation of the ictal EEG to the contralateral temporal lobe (P =0.001). Other historical data and interictal prose memory performance were not significantly correlated with absent aura. Ten of 11 patients without aura in monitoring also had absent or rare auras in their history. Conclusions: Lack of aura experience strongly correlates with indicators of bitemporal dysfunction such as bitemporal interictal sharp waves and bitemporal ictal propagation in scalp EEG, and absence of lateralized MRI sclerosis in patients with mesial TLE. The fact that absent auras are not correlated with episodic memory suggests a transient memory deficit, probably because of rapid propagation to the contralateral mesial temporal lobe. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Mesial temporal sclerosis; Temporal lobe epilepsy; Epilepsy; Memory; Aura
This study was in part presented at the 10th European Congress of Clinical Neurophysiology, Lyon, France, August 2000. * Corresponding author. Tel.: + 49-521-1444064; fax: + 49-521-1444562. E-mail address: [email protected] (R. Schulz).
0920-1211/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 9 5 - 9
202
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
1. Introduction: Auras in epileptic seizures are of diagnostic importance, indicating focality of the epileptic disorder as well as giving information for localization. The diagnostic value of the lack of aura experience is not clear. Penfield and Jasper (1954) were the first to report amnesia of a documented epileptic aura. However, they did not investigate this phenomenon in detail. Anterograde amnesia as failure to encode new semantic or episodic material is associated with bilateral mesial temporal lobe dysfunction (Scoville and Milner, 1957; Halgren et al., 1985; Squire 1986). Factors possibly indicating bilateral mesial temporal lobe dysfunction include: (1) bitemporal interictal epileptiform discharges (IED); (2) in invasive electroencephalography (EEG), rapid spread of the ictal EEG pattern to the contralateral temporal lobe; (3) in surface EEG, bilateral independent EEG seizure patterns or switch of lateralization of the EEG seizure pattern; (4) bilateral magnetic resonance imaging (MRI) pathology of the mesial temporal lobes; in neuropsychologic tests, (5) dysfunction of episodic memory; and (6) worse outcome following temporal lobe resection (Engel, 1994). A number of studies addressed various aspects of bilateral mesial temporal lobe dysfunction. In a previous study of patients with focal epilepsy, we found that amnesia of the aura was associated with seizure severity and rapid propagation of the ictal scalp EEG during the aura to the contralateral hemisphere. In that study, however, we did not investigate the further evolution of the ictal EEG following the aura, e.g. bilateral independent ictal scalp EEG propagation patterns or switch of lateralization (Schulz et al., 1995). Furthermore, we did not study patients who did not have auras in their history or in monitoring at all. In a study with bitemporal depth electrodes in temporal lobe epilepsy (TLE), Sperling et al. (1989) observed that patients with isolated auras were considered candidates for epilepsy surgery more often and had a better outcome than patients without isolated auras. This difference, however, did not reach statistical significance. The findings of Sperling et al. are consistent with those
reported by Steinhoff et al. (1995), who found auras significantly more often in the history of patients with unitemporal versus bitemporal IED in scalp EEG. In another recent study, we stated that worse outcome in epilepsy surgery of patients with mesial TLE is significantly correlated with independent scalp EEG seizure patterns over the contralateral temporal lobe, possibly indicating bitemporal epileptogenic zones (Schulz et al., 2000). Recently, Vargha-Khadem et al. (1997) described severe deficits of episodic memory but relatively intact semantic memory in three patients with bitemporal MRI pathology. Palmini et al. (1992) correlated pure amnestic seizures with bitemporal dysfunction in non-invasive and invasive (amobarbital) neuropsychologic memory tests. In this study, the predictive value of absence of aura experience in the patient’s history or during EEG/video monitoring on the results of presurgery diagnosis in mesial TLE will be evaluated. We analyze the correlation of aura history and experience of an aura during EEG/video monitoring with ictal scalp EEG propagation patterns, interictal EEG, prose memory performance, and presence of unilateral mesial temporal pathology in MRI. Our hypothesis was that absence of an aura in the patient’s history or during EEG/ video monitoring indicates bilateral mesial temporal lobe dysfunction.
2. Method In a prospective study over 3 years, 58 patients with medically refractory mesial TLE, because of mesial temporal lobe sclerosis, or with non-lesional mesial TLE were investigated with non-invasive EEG/video monitoring. The electrodes were placed according to the 10–10 system and, as a rule, sphenoidal electrodes were inserted. The clinical and demographic variables are summarized in Table 1. The location and frequency of IED were assessed by visual analysis of interictal EEG samples of 2 min duration every hour, including sleep. The evaluation of IED was supervised by one of the authors (R.S., A.E., or I.T.).
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
A total of 347 seizures were recorded (minimum three; range, 3–15; mean, 6.0 per patient).The ictal scalp EEG data was stored on digital tape, and was retrieved and printed out for analysis in 16-channel bipolar and referential montages by a laser printer. Pz was used as reference in referential montages that usually included the following electrodes: FP1, F3, C3, P3, O1, AF7, FC5, CP5, F7, FT7, T7, TP7, P7, SP1, F9, FT9, T9, TP9, homologous right-sided electrodes, PZ, CZ, FZ, and an extracranial electrode on the right shoulder (Ebner and Hoppe, 1995). Digital filtering and gain adjustment was used to optimize the EEG display. The laser printout of each seizure was analyzed by two board-certified electroencephalographers (R.S., German board; M.H., American and German board). One of the two electroencephalographers was blinded for all other data of the patients (M.H.). After independent review of all ictal EEGs for the assessment of interobserver reliability, agreement was reached on the categorization of all ictal EEGs that were evaluated differently by the two readers. Ictal EEGs tracings were categorized as follows: (1) regionalized right temporal or lateralized right hemisphere throughout the whole tracing; (2) regionalized left temporal or lateralized left hemisphere throughout the whole tracing; (3) non-lateralized; and (4) initially regionalized or lateralized ictal EEG onset: (a) followed by lateralization switched to the contralateral hemisphere Table 1 Clinical and demographic variables Patients Age Duration of epilepsy Precipitating injury (Mathern et al., 1995) Aura in patient’s history Aura in EEG/video monitoring
a
n= 58 8–57 years (mean, 30.8 years) 4–51 years (mean, 20.6 years) n= 36 (mean, 2.4 years) Yes, n= 51; no, n= 7 Yes, n = 47; no, n = 11a
n= 11 comprises seven patients who had no aura in their history and four patients who had no aura in monitoring but had rare auras in their history.
203
or the contralateral temporal lobe, or (b) followed by a bilateral asynchronous ictal EEG pattern with a frequency difference of 1 or \ 1 Hz between the two hemispheres or the two temporal lobes. An ictal pattern was defined as a regionalized temporal ictal EEG when the amplitude in one temporal chain was at least twice as high as the amplitude in the ipsilateral parasagittal bipolar montages. An ictal pattern was defined as lateralized if the amplitude in the referential montages of one hemisphere was at least twice the amplitude in the contralateral hemisphere (Steinhoff et al., 1995). If Pz was active, an ictal EEG was classified as lateralized or regionalized when the amplitude in the bipolar chain of one hemisphere was at least twice the amplitude of the bipolar chain in the contralateral hemisphere. According to current practice, delayed regionalization within 30 s was judged equivalent to initial regionalization (Risinger et al., 1989). On the other hand, the term of ‘regionalized EEG seizure pattern’ was used differently to current practice because, up to now, propagation after regionalization has rarely been considered in the literature (Steinhoff et al., 1995; Schulz et al., 2000). The diagnosis of switch of lateralization or asynchrony of EEG seizure pattern was only made if a regionalized or lateralized pattern lasted for at least 10 s. An example of an EEG seizure pattern with switch of lateralization and asynchrony is shown in Schulz et al. (2000). After categorizing the ictal EEGs of all patients according to the criteria already outlined, patients were classified as follows: (1) patients in whom all ictal EEGs were either regionalized right temporal/right hemispheric or left temporal/left hemispheric (one non-lateralized EEG was also allowed); (2) patients in whom all or most ictal EEGs were non-lateralized (the other ictal EEGs could be regionalized or lateralized); (3) patients in whom a switch of lateralization or asynchronous EEG pattern occurred during at least one seizure; and (4) patients in whom, in some seizures, lateralization occurred to one hemisphere or temporal lobe and, in other seizures, to the contralateral hemisphere or temporal lobe. Patients that had both independent EEG seizure patterns in the right and left temporal lobe and
204
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
also showed switch of lateralization or asynchronous EEG patterns were assigned to patient category (4). In all 58 patients, a detailed history was taken including onset of seizures, major events during pregnancy and delivery, development, presence of an aura within the past years or in earlier years of epilepsy, and frequency of seizures. During monitoring, recall or not of an aura following a seizure was documented after complete postictal recovery of orientation in all 347 seizures. Theoretically, the absence of an aura in the postictal interview could result from the absence of an aura or from the failure to adequately encode an aura preceding the seizure. In most patients, antiepileptic medication was reduced to facilitate the occurrence of seizures. Seizures were classified according to their severity in the following three categories: (1) psychomotor seizures with a duration of automatisms or loss of consciousness of 30 s or less; (2) psychomotor seizures with a duration of automatisms or loss of consciousness of more than 30 s; and (3) secondarily generalized tonic clonic seizures. In this manuscript, psychomotor seizures were defined as partial seizures with motor automatisms usually associated with loss of responsiveness. Regarding the presence or absence of aura memory in EEG/video monitoring, all patients were categorized as follows: group 1, 14 patients with memory of an aura after each seizure; group 2, 33 patients with memory of an aura after some of the seizures; and group 3, 11 patients with no memory of an aura after all seizures, and also no occurrence of isolated auras during monitoring. The latter group of 11 patients, with no memory of an aura after all seizures, consisted of two subgroups: group 3a, seven patients who also did not recall having auras historically; and group 3b, four patients who recalled having auras only rarely in their history (in two patients, respectively, 20 and 7 years before the EEG/video monitoring; in one patient, only with some of the seizures; another patient could not estimate the frequency). The four patients of group 3b with rare auras in their histories had 18 recorded seizures without recalling an aura. Of these 18 recorded seizures, two were secondarily general-
ized tonic clonic seizures, 14 were psychomotor seizures with a duration of more than 30 s, and two were psychomotor seizures with a duration of less than 30 s. Isolated auras were not recorded. The two generalized tonic clonic seizures occurred in a patient who, in addition, had four psychomotor seizures of more than 30 s duration. There were no patients who had no aura in history but who had auras in monitoring. Fifty-three patients had an extensive neuropsychologic evaluation. The age-adjusted prose memory percentile from the Wechsler Memory Scale Revised (WSM-R) (logical memory) was used as an index of possible interictal memory deficits that could be relevant to explain the amnesia for the epileptic aura. In four children and in three adults, quantitative WSM-R testing was not possible. In these patients, the memory function was estimated from previous neuropsychologic tests or from the subjective impression of the neuropsychologist who delivered the testing battery. In addition, in one child who did not speak German and in four adults, no reliable information could be gathered. High-resolution MRIs were obtained with a Siemens Magnetom Impact 1.0-T scanner, and included T1-weighted three-dimensional volume, proton-density, and T2-weighted images. All MRIs were evaluated by one experienced investigator (R.S.). For visual evaluation, coronal 3 mm T1-weighted three-dimensional volume slices and coronal 3 mm T2-weighted slices were used. In some patients, coronal inversion recovery (IR) images were also available. The MRI were classified in the following categories: (1) absent, (2) mild, and (3) marked unilateral hippocampal atrophy or, in IR images, clear loss of signal or, in T2 images, increase of signal of mesial temporal structures. A Fluorodeoxy Glucose Position Emission Tomography study was carried out in 45 patients. Some of the patients in whom no PET study was performed were considered good surgical candidates with sufficient concordant information from other methods. The other patients without a PET study were not considered surgical candidates because of contradictory results obtained with the other methods.
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
205
Fig. 1. EEG seizure propagation (in percentage of patients) as a function of postictal memory of an aura in EEG/video monitoring. SOL, Switch of lateralization.
The diagnoses of TLE and lateralization of TLE were based on MRI results, interictal and ictal EEG, PET studies, and clinical semiology. Lateralization and localization of the epilepsy was assumed to be unreliable when none or only one of the following factors were present: (1) clear lateralized mesial temporal pathology in MRI; (2) 100% IED at one temporal lobe; (3) all EEG seizure patterns regionalized or lateralized to one temporal lobe without later switch of lateralization or asynchrony (one non-lateralized EEG seizure was allowed); and (4) significant unitemporal hypometabolism in PET (15% or more in two or more slices). Forty-nine patients had a temporal lobe resection. The resection included the mesial temporal structures (amygdala, hippocampus, parahippocampus), and parts of the anterior, lateral and basal temporal neocortex (keyhole or standard temporal lobe resection). Histology confirmed the diagnosis of mesial temporal sclerosis in all 49 cases (in two out of 49, only end-folium gliosis was found).
For statistical analysis, the following tests were carried out with the SPSS statistical package, version 7.5 for Windows (SPSS Inc., Chicago, IL): exact Fisher test, chi-square test for homogeneity or the test of trend, and the Mann –Whitney Test (U-test). The interobserver reliability was assessed according to Cyr and Francis (1992).
3. Results Ictal scalp EEG propagation to the contralateral temporal lobe, bitemporal IED, and absence of unilateral mesial temporal pathology in MRI were significantly correlated with the absence of aura experience during monitoring and absent or rare auras in history. Fig. 1 shows the ictal scalp EEG propagation patterns as a function of postictal memory of auras. The number of EEG seizure patterns that remained regionalized or lateralized to one hemisphere decreased significantly from 71.4% (ten out of 14) in patients who always recalled an aura, to 60.6% (20 out of 33) in patients who sometimes
206
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
recalled an aura, and to 18.2% (two out of 11) in patients who never recalled an aura and had no isolated aura during monitoring (chi-square test of trend, P=0.001). Fig. 2 shows the distribution of IED as a function of the memory of auras. The percentage of patients with 100% IED at one temporal lobe decreased significantly from 71.4% (ten out of 14) in patients who always recalled an aura, to 54.5% (18 of 33) in patients who sometimes recalled an aura, to 18.2% (two out of 11) in patients who never recalled an aura (chi-square test of trend, P =0.008). Fig. 3 shows the distribution of lateralized mesial temporal MRI pathology as a function of the memory of auras. The percentage of patients without clear unitemporal pathology was 7.1% (one out of 14) in patients who always recalled an aura, 9.4% (three out of 32) in patients who sometimes recalled an aura, and 36.4% (four out of 11) in patients who had no aura (Fisher’s exact test, two-sided, P =0.004). MRI was not available for re-evaluation in one patient. There was no correlation of marked or moderate unilateral
mesial temporal MRI pathology regarding absence of an aura. Lateralized mesial temporal MRI pathology and ictal scalp EEG propagation patterns were not significantly correlated (Fisher’s exact test, two-sided, P= 0.402). Four of eight patients without unilateral mesial temporal MRI pathology showed bitemporal asynchronous ictal EEG or switch of lateralization, and the remaining four patients had a regionalized ictal EEG. Out of the 49 patients with unilateral mesial temporal MRI pathology, two had independent left and right ictal EEG in different seizures, 11 had bitemporal asynchronous ictal EEG or switch of lateralization, nine patients had a non-lateralized ictal EEG, and 27 patients had a regionalized ictal EEG. By visual analysis, one of the eight patients without lateralized mesial temporal MRI pathology had bilateral mesial temporal lobe atrophy in MRI; this patient, with a history of encephalitis, had bitemporal IED and a left temporal ictal EEG. No significant correlations could be shown between recollection of an aura and duration of
Fig. 2. Interictal epileptiform discharges (IED) as a function of postictal memory of an aura in EEG/video monitoring.
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
207
Fig. 3. Unilateral mesial temporal pathology in MRI as a function of postictal memory of an aura in EEG/video monitoring.
epilepsy, prose memory performance, and age of occurrence of possible etiological factor (Mann – Whitney test, P [ 0.328). Aura memory in video/EEG monitoring was also a function of the reliability of the localization and lateralization of the mesial TLE. Failure to recall an aura after each seizure occurred significantly more often in patients with unreliably diagnosed lateralization of mesial TLE (54.5%, six of 11 patients) than in patients with reliably diagnosed lateralization of mesial TLE (right TLE, 4.8% (one of 21 patients); left TLE, 15.4% (four of 26 patients); Fisher’s exact test, two-sided, P= 0.031). Significant differences between right and left TLE could not be found with regard to aura memory. The diagnosis of temporal lobe epilepsy as opposed to extratemporal epilepsy could be confirmed in all eight patients without lateralized MRI pathology: four of the eight patients had significantly lateralized temporal hypometabolism, and three had non-lateralized temporal hypometabolism; the remaining patient had been studied with subdural and depth electrodes in another epilepsy center, with the result of independent right and left temporal subclinical and clinical EEG seizure patterns.
4. Discussion This study shows that patients with mesial TLE, without or with only rare auras (group 3 patients), had signs of bilateral mesial temporal lobe dysfunction. Evidence is given by the following findings: (1) interictal scalp EEG (bitemporal IED); (2) ictal scalp EEG (bilateral independent, asynchronous EEG seizure patterns or switch of lateralization); and (3) the absence of lateralized MRI pathology of the mesial temporal lobes. The correlation of group 3 patients with bitemporal IED is consistent with Steinhoff’s findings of a significantly rarer aura in patients with bitemporal versus unitemporal IED in scalp EEG (Steinhoff et al., 1995). With regard to the ictal EEG, possible reasons for not having auras or not recalling auras could be as follows. (1) Rapid propagation of the epileptiform discharge to the contralateral temporal lobe with bilateral inactivation of the mesial temporal structures that are essential for memory encoding. This could be related to epileptogenicity of the contralateral temporal lobe or relative severity of the seizures. Both mechanisms would facilitate bilateral hippocampal seizure spread. (2)
208
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
Epileptic discharge, inactivating one mesial temporal structure with morphological damage to the contralateral hippocampus (this would also lead to functional inactivation of both mesial temporal structures). (3) Rapid propagation of the epileptic discharge into both temporal lobes from an extratemporal epileptogenic zone, e.g. the parietal or occipital lobes. Our findings of bilateral, independent, asynchronous EEG seizure patterns or switch of lateralization of the EEG seizure pattern in surface EEG in group 3 patients corroborate a previous study of a heterogeneous group of patients with focal epilepsy (Schulz et al., 1995). In this study, we found a correlation of ictal scalp EEG propagation to the contralateral hemisphere with amnesia of the previously documented aura, but we had addressed individual seizures and had considered the ictal EEG only during the aura. Furthermore, we did not differentiate asynchronous bitemporal ictal EEG patterns and switch of lateralization from EEG seizure patterns with identical frequency over both hemispheres so that a bilateral EEG seizure pattern during the aura could have resulted from volume conduction to the contralateral anterior and mesial (sphenoidal) electrodes or from a diffuse EEG seizure pattern with later regionalization according to the principles stated by Risinger et al. (1989). The present study defines the EEG patterns of switch of lateralization and bilateral asynchrony, both of which can be regarded as an independent EEG seizure pattern involving the contralateral temporal lobe in patients with mesial TLE (Schulz et al., 2000). Severity of the seizure was not a likely cause for failure to recollect an aura in group 3 patients because secondarily generalized tonic – clonic seizures were also rare in this patient group. Actually, the only patient in this group who had secondary generalization also had four additional psychomotor seizures without recollection of an aura. On the other hand, a switch of lateralization of the EEG seizure and an asynchronous EEG seizure pattern over both temporal lobes were found significantly more often in patients without aura, consistent with our hypothesis of a bitemporal dysfunction.
A discrete lesion of the contralateral temporal lobe or secondary contralateral epileptogenesis are possible causes of bitemporal dysfunction. Morrell (1989) established the concept of secondary epileptogenesis by stimulation studies in animals and by studies of epileptic patients with unilateral tumors. In this concept, frequent ictal propagation to the contralateral temporal lobe is thought to cause contralateral IED. Bilateral independent EEG seizure patterns could also result from a more extensive damage to the ipsilateral mesial temporal lobe, which might facilitate faster contralateral propagation. In this case, patients with marked mesial temporal sclerosis on MRI should have bitemporal ictal patterns more often than patients without or with only moderate unilateral mesial pathology. However, we found no correlation of independent bitemporal ictal propagation with unilateral MRI pathology. Therefore, independent epileptogenicity of the contralateral temporal lobe or a different mode of propagation across the frontal lobes (Lieb et al., 1991) or the posterior hippocampal commissure (Gloor et al., 1993) might explain the absence of auras in this group. Propagation from an extratemporal epileptogenic zone into both temporal lobes is difficult to exclude in non-lesional epilepsies. Temporal hypometabolism in PET in seven of eight patients without unilateral mesial temporal MRI pathology and investigation with subdural and depth electrodes in the remaining patient make an extratemporal epileptogenic zone less probable. Bitemporal dysfunction can result in poor memory functions in general. Vargha-Khadem et al. (1997) described severe deficits of episodic memory in three patients with marked bitemporal MRI pathology. In our study, aura history and aura recall were not significantly related with the recall of prose passages as an indicator of episodic memory. Group 3 patients significantly more often had bitemporal IED and bitemporal independent ictal EEG; but they also significantly more often had no unilateral mesial temporal MRI pathology. With the exception of one patient, we did not find marked bitemporal pathology in this group, which would have been detectable by visual analysis similar to the patients of Vargha-
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
Khadem et al. We suppose that more discrete bitemporal pathology that is not detectable by visual analysis causes no severe interictal deficits of episodic memory, but only the comparatively discrete signs of amnesia of an aura or failure to experience an aura. According to our alternative hypothesis, the interictal absence of severe deficits of episodic memory in Group 3 patients could also be explained as an exclusively ictal phenomenon by rapid propagation and inactivation of the contralateral mesial temporal lobe. MRI volumetry of both hippocampi, with T2 relaxometry, and with MR spectroscopy could be of interest to assess discrete bitemporal mesial pathology. It is also theoretically possible that patients without aura during monitoring still suffer from transient postictal memory deficits. Bergin et al. (1995) found no memory deficits during monitoring after 48 h in spite of complex partial seizures, secondarily generalized tonic – clonic seizures or both in the interval. Helmstaedter et al. (1994) described verbal and non-verbal learning deficits with a duration of 30 min to 1 h after complete reorientation. Helmstaedter et al. tested postictal learning but not postictal retrieval of preictally encoded episodic material. We think that reorientation with respect to person, location and time represents intact episodic retrieval so that the failure to recall an aura postictally should reliably indicate the absence of an aura or the failure to encode an aura at the beginning of the seizure. All our patients were asked for memory of an aura after renewed postictal reactivity, reorientation and resolved possible postictal aphasia. Postictal memory deficits in some patients because of insufficient encoding could indicate bitemporal mesial dysfunction according to our hypothesis. The relatively poor surgical outcome in patients with switch of lateralization and asynchronous EEG seizures (Schulz et al., 2000) also supports the hypothesis of bitemporal epileptogenicity in group 3 patients. Another explanation for a worse prognosis could be a more posterior pathology in these patients together with an incomplete resection of the most posterior parts of the hippocampus and parahippocampus, so that propagation from remaining epileptogenic tissue across the
209
posterior hippocampal commissure might occur. On the other hand, postoperative remains of epileptogenic tissue are more probable in patients with mesial temporal sclerosis visible on MRI than in patients without visible unilateral mesial pathology. Our study shows that patients with mesial TLE without or with only rare auras in their history, who had no auras during non-invasive EEG/video monitoring, significantly more often had bitemporal IED, ictal patterns suggestive of a bitemporal independent epileptogenicity, absence of lateralized mesial temporal MRI pathology, and a worse outcome, which together indicate a bilateral mesial temporal disease or a different seizure propagation mode. Patients with the aforementioned clinical features might therefore form a subgroup of patients with mesial TLE who often need invasive monitoring with subdural strips or depth electrodes.
Acknowledgements The authors thank Ralf Dernbach, Martina Fru¨nd, and their teams of EEG technicians and nurses for their commitment in the ictal and postictal testing of the patients.
References Bergin, P.S., Thompson, P.J., Fish, D.R., Shorvon, S.D., 1995. The effect of seizures on memory for recently learned material. Neurology 45, 236– 240. Cyr, L., Francis, K., 1992. Measures of clinical agreement for nominal and categorical data. The kappa coefficient. Comput. Biol. Med. 22, 239– 246. Ebner, A., Hoppe, M., 1995. Noninvasive electroencephalography and mesial temporal sclerosis. J. Clin. Neurophysiol. 12, 23 – 31. Engel, J., 1994. Bilateral temporal lobe epilepsy. In: Wolf, P. (Ed.), Epileptic Seizures and Syndromes. John Libbey, London, pp. 359– 368. Gloor, P., Salanova, V., Olivier, A., Quesney, L.F., 1993. The dorsal hippocampal commissure. An anatomically identifiable and functional pathway. Brain 116, 1249– 1273. Halgren, E., Wilson, C.L., Stapleton, J.M., 1985. Human medial temporallobe stimulation disrupts both formation and retrieval of recent memories. Brain Cogn. 4, 287– 295.
210
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210 Schulz, R., Lu¨ders, H.O., Noachtar, S., May, T., Sakamoto, A., Holhausen, H., Wolf, P., 1995. Amnesia of the epileptic aura. Neurology 45, 231– 235. Schulz, R., Lu¨ders, H.O., Hoppe, M., Tuxhorn, I., May, T., Ebner, A., 2000. Interictal EEG and ictal scalp EEG propagation are highly predictive of surgical outcome in mesial temporal lobe epilepsy. Epilepsia 41, 564– 570. Scoville, W.B., Milner, B., 1957. Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatr. 20, 11 – 21. Sperling, M.R., Lieb, J.P., Engel, J., Crandall, P.H., 1989. Prognostic significance of independent auras in temporal lobe seizures. Epilepsia 30, 322– 331. Squire, L.R., 1986. Mechanisms of memory. Science 232, 1612– 1619. Steinhoff, B.J., So, N.K., Lim, S., Lu¨ders, H.O., 1995. Ictal scalp EEG in temporal lobe epilepsy with unitemporal versus bitemporal interictal epileptiform discharges. Neurology 45, 889– 896. Vargha-Khadem, F., Gadian, D.G., Watkins, K.E., Connelly, A., Van Paesschen, W., Mishkin, M., 1997. Differential effects of early hippocampal pathology on episodic and semantic memory. Science 277, 376– 380.
Helmstaedter, C., Elger, C.E., Lendt, M., 1994. Postictal courses of cognitive deficits in focal epilepsies. Epilepsia 35, 1073– 1078. Lieb, J.P., Dasheiff, R.M., Engel, J., 1991. Role of the frontal lobes in the propagation of mesial temporal lobe seizures. Epilepsia 32, 822– 837. Mathern, G.W., Babb, T.L., Vickrey, B.G., Melendez, M., Pretorius, J.K., 1995. The clinical–pathogenetic mechanisms of hippocampal neuron loss and surgical outcomes in temporal lobe epilepsy. Brain 118, 105–118. Morrell, F., 1989. Varieties of human secondary epileptogenesis. J. Clin. Neurophysiol. 6, 227–275. Palmini, A.L., Gloor, P., Jones-Gotman, M., 1992. Pure amnestic seizures in temporal lobe epilepsy. Brain 115, 749– 769. Penfield, W., Jasper, H., 1954. Epilepsy and the Functional Anatomy of the Human Brain. Little, Brown & Co, Boston, MA, p. 528. Risinger, M.W., Engel, J., Van Ness, P.C., Henry, T.R., Crandall, P.H., 1989. Ictal localization of temporal lobe seizures with scalp/ sphenoidal recordings. Neurology 39, 1288– 1293.
.
Lack of aura experience correlates with bitemporal dysfunction in mesial temporal lobe epilepsy Reinhard Schulz a,*, Hans O. Lu¨ders b, Matthias Hoppe a, Hennric Jokeit a, Anke Moch a, Ingrid Tuxhorn a, Theodor May c, Alois Ebner a a
Abteilung fu¨r pra¨chirurgische Epilepsiediagnostik, Epilepsiezentrum Mara gGmbH, Epilepsiezentrum Bethel, Maraweg 21, 33617 Bielefeld, Germany b Cle6eland Clinic Foundation, Cle6eland, OH, USA c Gesellschaft fu¨r Epilepsieforschung, 33617 Bielefeld, Germany Received 27 July 2000; received in revised form 5 October 2000; accepted 15 October 2000
Abstract The diagnostic value of lack of aura experience in patients with temporal lobe epilepsy (TLE) is unclear. Purpose: To evaluate possible factors of bitemporal dysfunction in patients with mesial TLE who did not experience an aura in electroencephalography EEG/video monitoring for epilepsy surgery. Methods: Ictal scalp EEG propagation patterns of 347 seizures of 58 patients with mesial temporal lobe sclerosis or non-lesional mesial TLE, interictal epileptiform discharges (IED), presence of unilateral mesial temporal lobe sclerosis in visual magnetic resonance imaging (MRI) analysis, prose memory performance, history or not of an aura, and postictal memory or absence of an aura were analyzed. The ictal EEG was categorized as follows. EEG seizure: (a) remaining regionalized, (b) non-lateralized, (c) showing later switch of lateralization or bitemporal asynchronous ictal patterns. Results: Absent aura in monitoring was significantly correlated with absence of unitemporal MRI sclerosis (P= 0.004), bitemporal IED (P = 0.008), and propagation of the ictal EEG to the contralateral temporal lobe (P =0.001). Other historical data and interictal prose memory performance were not significantly correlated with absent aura. Ten of 11 patients without aura in monitoring also had absent or rare auras in their history. Conclusions: Lack of aura experience strongly correlates with indicators of bitemporal dysfunction such as bitemporal interictal sharp waves and bitemporal ictal propagation in scalp EEG, and absence of lateralized MRI sclerosis in patients with mesial TLE. The fact that absent auras are not correlated with episodic memory suggests a transient memory deficit, probably because of rapid propagation to the contralateral mesial temporal lobe. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Mesial temporal sclerosis; Temporal lobe epilepsy; Epilepsy; Memory; Aura
This study was in part presented at the 10th European Congress of Clinical Neurophysiology, Lyon, France, August 2000. * Corresponding author. Tel.: + 49-521-1444064; fax: + 49-521-1444562. E-mail address: [email protected] (R. Schulz).
0920-1211/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 9 5 - 9
202
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
1. Introduction: Auras in epileptic seizures are of diagnostic importance, indicating focality of the epileptic disorder as well as giving information for localization. The diagnostic value of the lack of aura experience is not clear. Penfield and Jasper (1954) were the first to report amnesia of a documented epileptic aura. However, they did not investigate this phenomenon in detail. Anterograde amnesia as failure to encode new semantic or episodic material is associated with bilateral mesial temporal lobe dysfunction (Scoville and Milner, 1957; Halgren et al., 1985; Squire 1986). Factors possibly indicating bilateral mesial temporal lobe dysfunction include: (1) bitemporal interictal epileptiform discharges (IED); (2) in invasive electroencephalography (EEG), rapid spread of the ictal EEG pattern to the contralateral temporal lobe; (3) in surface EEG, bilateral independent EEG seizure patterns or switch of lateralization of the EEG seizure pattern; (4) bilateral magnetic resonance imaging (MRI) pathology of the mesial temporal lobes; in neuropsychologic tests, (5) dysfunction of episodic memory; and (6) worse outcome following temporal lobe resection (Engel, 1994). A number of studies addressed various aspects of bilateral mesial temporal lobe dysfunction. In a previous study of patients with focal epilepsy, we found that amnesia of the aura was associated with seizure severity and rapid propagation of the ictal scalp EEG during the aura to the contralateral hemisphere. In that study, however, we did not investigate the further evolution of the ictal EEG following the aura, e.g. bilateral independent ictal scalp EEG propagation patterns or switch of lateralization (Schulz et al., 1995). Furthermore, we did not study patients who did not have auras in their history or in monitoring at all. In a study with bitemporal depth electrodes in temporal lobe epilepsy (TLE), Sperling et al. (1989) observed that patients with isolated auras were considered candidates for epilepsy surgery more often and had a better outcome than patients without isolated auras. This difference, however, did not reach statistical significance. The findings of Sperling et al. are consistent with those
reported by Steinhoff et al. (1995), who found auras significantly more often in the history of patients with unitemporal versus bitemporal IED in scalp EEG. In another recent study, we stated that worse outcome in epilepsy surgery of patients with mesial TLE is significantly correlated with independent scalp EEG seizure patterns over the contralateral temporal lobe, possibly indicating bitemporal epileptogenic zones (Schulz et al., 2000). Recently, Vargha-Khadem et al. (1997) described severe deficits of episodic memory but relatively intact semantic memory in three patients with bitemporal MRI pathology. Palmini et al. (1992) correlated pure amnestic seizures with bitemporal dysfunction in non-invasive and invasive (amobarbital) neuropsychologic memory tests. In this study, the predictive value of absence of aura experience in the patient’s history or during EEG/video monitoring on the results of presurgery diagnosis in mesial TLE will be evaluated. We analyze the correlation of aura history and experience of an aura during EEG/video monitoring with ictal scalp EEG propagation patterns, interictal EEG, prose memory performance, and presence of unilateral mesial temporal pathology in MRI. Our hypothesis was that absence of an aura in the patient’s history or during EEG/ video monitoring indicates bilateral mesial temporal lobe dysfunction.
2. Method In a prospective study over 3 years, 58 patients with medically refractory mesial TLE, because of mesial temporal lobe sclerosis, or with non-lesional mesial TLE were investigated with non-invasive EEG/video monitoring. The electrodes were placed according to the 10–10 system and, as a rule, sphenoidal electrodes were inserted. The clinical and demographic variables are summarized in Table 1. The location and frequency of IED were assessed by visual analysis of interictal EEG samples of 2 min duration every hour, including sleep. The evaluation of IED was supervised by one of the authors (R.S., A.E., or I.T.).
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
A total of 347 seizures were recorded (minimum three; range, 3–15; mean, 6.0 per patient).The ictal scalp EEG data was stored on digital tape, and was retrieved and printed out for analysis in 16-channel bipolar and referential montages by a laser printer. Pz was used as reference in referential montages that usually included the following electrodes: FP1, F3, C3, P3, O1, AF7, FC5, CP5, F7, FT7, T7, TP7, P7, SP1, F9, FT9, T9, TP9, homologous right-sided electrodes, PZ, CZ, FZ, and an extracranial electrode on the right shoulder (Ebner and Hoppe, 1995). Digital filtering and gain adjustment was used to optimize the EEG display. The laser printout of each seizure was analyzed by two board-certified electroencephalographers (R.S., German board; M.H., American and German board). One of the two electroencephalographers was blinded for all other data of the patients (M.H.). After independent review of all ictal EEGs for the assessment of interobserver reliability, agreement was reached on the categorization of all ictal EEGs that were evaluated differently by the two readers. Ictal EEGs tracings were categorized as follows: (1) regionalized right temporal or lateralized right hemisphere throughout the whole tracing; (2) regionalized left temporal or lateralized left hemisphere throughout the whole tracing; (3) non-lateralized; and (4) initially regionalized or lateralized ictal EEG onset: (a) followed by lateralization switched to the contralateral hemisphere Table 1 Clinical and demographic variables Patients Age Duration of epilepsy Precipitating injury (Mathern et al., 1995) Aura in patient’s history Aura in EEG/video monitoring
a
n= 58 8–57 years (mean, 30.8 years) 4–51 years (mean, 20.6 years) n= 36 (mean, 2.4 years) Yes, n= 51; no, n= 7 Yes, n = 47; no, n = 11a
n= 11 comprises seven patients who had no aura in their history and four patients who had no aura in monitoring but had rare auras in their history.
203
or the contralateral temporal lobe, or (b) followed by a bilateral asynchronous ictal EEG pattern with a frequency difference of 1 or \ 1 Hz between the two hemispheres or the two temporal lobes. An ictal pattern was defined as a regionalized temporal ictal EEG when the amplitude in one temporal chain was at least twice as high as the amplitude in the ipsilateral parasagittal bipolar montages. An ictal pattern was defined as lateralized if the amplitude in the referential montages of one hemisphere was at least twice the amplitude in the contralateral hemisphere (Steinhoff et al., 1995). If Pz was active, an ictal EEG was classified as lateralized or regionalized when the amplitude in the bipolar chain of one hemisphere was at least twice the amplitude of the bipolar chain in the contralateral hemisphere. According to current practice, delayed regionalization within 30 s was judged equivalent to initial regionalization (Risinger et al., 1989). On the other hand, the term of ‘regionalized EEG seizure pattern’ was used differently to current practice because, up to now, propagation after regionalization has rarely been considered in the literature (Steinhoff et al., 1995; Schulz et al., 2000). The diagnosis of switch of lateralization or asynchrony of EEG seizure pattern was only made if a regionalized or lateralized pattern lasted for at least 10 s. An example of an EEG seizure pattern with switch of lateralization and asynchrony is shown in Schulz et al. (2000). After categorizing the ictal EEGs of all patients according to the criteria already outlined, patients were classified as follows: (1) patients in whom all ictal EEGs were either regionalized right temporal/right hemispheric or left temporal/left hemispheric (one non-lateralized EEG was also allowed); (2) patients in whom all or most ictal EEGs were non-lateralized (the other ictal EEGs could be regionalized or lateralized); (3) patients in whom a switch of lateralization or asynchronous EEG pattern occurred during at least one seizure; and (4) patients in whom, in some seizures, lateralization occurred to one hemisphere or temporal lobe and, in other seizures, to the contralateral hemisphere or temporal lobe. Patients that had both independent EEG seizure patterns in the right and left temporal lobe and
204
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
also showed switch of lateralization or asynchronous EEG patterns were assigned to patient category (4). In all 58 patients, a detailed history was taken including onset of seizures, major events during pregnancy and delivery, development, presence of an aura within the past years or in earlier years of epilepsy, and frequency of seizures. During monitoring, recall or not of an aura following a seizure was documented after complete postictal recovery of orientation in all 347 seizures. Theoretically, the absence of an aura in the postictal interview could result from the absence of an aura or from the failure to adequately encode an aura preceding the seizure. In most patients, antiepileptic medication was reduced to facilitate the occurrence of seizures. Seizures were classified according to their severity in the following three categories: (1) psychomotor seizures with a duration of automatisms or loss of consciousness of 30 s or less; (2) psychomotor seizures with a duration of automatisms or loss of consciousness of more than 30 s; and (3) secondarily generalized tonic clonic seizures. In this manuscript, psychomotor seizures were defined as partial seizures with motor automatisms usually associated with loss of responsiveness. Regarding the presence or absence of aura memory in EEG/video monitoring, all patients were categorized as follows: group 1, 14 patients with memory of an aura after each seizure; group 2, 33 patients with memory of an aura after some of the seizures; and group 3, 11 patients with no memory of an aura after all seizures, and also no occurrence of isolated auras during monitoring. The latter group of 11 patients, with no memory of an aura after all seizures, consisted of two subgroups: group 3a, seven patients who also did not recall having auras historically; and group 3b, four patients who recalled having auras only rarely in their history (in two patients, respectively, 20 and 7 years before the EEG/video monitoring; in one patient, only with some of the seizures; another patient could not estimate the frequency). The four patients of group 3b with rare auras in their histories had 18 recorded seizures without recalling an aura. Of these 18 recorded seizures, two were secondarily general-
ized tonic clonic seizures, 14 were psychomotor seizures with a duration of more than 30 s, and two were psychomotor seizures with a duration of less than 30 s. Isolated auras were not recorded. The two generalized tonic clonic seizures occurred in a patient who, in addition, had four psychomotor seizures of more than 30 s duration. There were no patients who had no aura in history but who had auras in monitoring. Fifty-three patients had an extensive neuropsychologic evaluation. The age-adjusted prose memory percentile from the Wechsler Memory Scale Revised (WSM-R) (logical memory) was used as an index of possible interictal memory deficits that could be relevant to explain the amnesia for the epileptic aura. In four children and in three adults, quantitative WSM-R testing was not possible. In these patients, the memory function was estimated from previous neuropsychologic tests or from the subjective impression of the neuropsychologist who delivered the testing battery. In addition, in one child who did not speak German and in four adults, no reliable information could be gathered. High-resolution MRIs were obtained with a Siemens Magnetom Impact 1.0-T scanner, and included T1-weighted three-dimensional volume, proton-density, and T2-weighted images. All MRIs were evaluated by one experienced investigator (R.S.). For visual evaluation, coronal 3 mm T1-weighted three-dimensional volume slices and coronal 3 mm T2-weighted slices were used. In some patients, coronal inversion recovery (IR) images were also available. The MRI were classified in the following categories: (1) absent, (2) mild, and (3) marked unilateral hippocampal atrophy or, in IR images, clear loss of signal or, in T2 images, increase of signal of mesial temporal structures. A Fluorodeoxy Glucose Position Emission Tomography study was carried out in 45 patients. Some of the patients in whom no PET study was performed were considered good surgical candidates with sufficient concordant information from other methods. The other patients without a PET study were not considered surgical candidates because of contradictory results obtained with the other methods.
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
205
Fig. 1. EEG seizure propagation (in percentage of patients) as a function of postictal memory of an aura in EEG/video monitoring. SOL, Switch of lateralization.
The diagnoses of TLE and lateralization of TLE were based on MRI results, interictal and ictal EEG, PET studies, and clinical semiology. Lateralization and localization of the epilepsy was assumed to be unreliable when none or only one of the following factors were present: (1) clear lateralized mesial temporal pathology in MRI; (2) 100% IED at one temporal lobe; (3) all EEG seizure patterns regionalized or lateralized to one temporal lobe without later switch of lateralization or asynchrony (one non-lateralized EEG seizure was allowed); and (4) significant unitemporal hypometabolism in PET (15% or more in two or more slices). Forty-nine patients had a temporal lobe resection. The resection included the mesial temporal structures (amygdala, hippocampus, parahippocampus), and parts of the anterior, lateral and basal temporal neocortex (keyhole or standard temporal lobe resection). Histology confirmed the diagnosis of mesial temporal sclerosis in all 49 cases (in two out of 49, only end-folium gliosis was found).
For statistical analysis, the following tests were carried out with the SPSS statistical package, version 7.5 for Windows (SPSS Inc., Chicago, IL): exact Fisher test, chi-square test for homogeneity or the test of trend, and the Mann –Whitney Test (U-test). The interobserver reliability was assessed according to Cyr and Francis (1992).
3. Results Ictal scalp EEG propagation to the contralateral temporal lobe, bitemporal IED, and absence of unilateral mesial temporal pathology in MRI were significantly correlated with the absence of aura experience during monitoring and absent or rare auras in history. Fig. 1 shows the ictal scalp EEG propagation patterns as a function of postictal memory of auras. The number of EEG seizure patterns that remained regionalized or lateralized to one hemisphere decreased significantly from 71.4% (ten out of 14) in patients who always recalled an aura, to 60.6% (20 out of 33) in patients who sometimes
206
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
recalled an aura, and to 18.2% (two out of 11) in patients who never recalled an aura and had no isolated aura during monitoring (chi-square test of trend, P=0.001). Fig. 2 shows the distribution of IED as a function of the memory of auras. The percentage of patients with 100% IED at one temporal lobe decreased significantly from 71.4% (ten out of 14) in patients who always recalled an aura, to 54.5% (18 of 33) in patients who sometimes recalled an aura, to 18.2% (two out of 11) in patients who never recalled an aura (chi-square test of trend, P =0.008). Fig. 3 shows the distribution of lateralized mesial temporal MRI pathology as a function of the memory of auras. The percentage of patients without clear unitemporal pathology was 7.1% (one out of 14) in patients who always recalled an aura, 9.4% (three out of 32) in patients who sometimes recalled an aura, and 36.4% (four out of 11) in patients who had no aura (Fisher’s exact test, two-sided, P =0.004). MRI was not available for re-evaluation in one patient. There was no correlation of marked or moderate unilateral
mesial temporal MRI pathology regarding absence of an aura. Lateralized mesial temporal MRI pathology and ictal scalp EEG propagation patterns were not significantly correlated (Fisher’s exact test, two-sided, P= 0.402). Four of eight patients without unilateral mesial temporal MRI pathology showed bitemporal asynchronous ictal EEG or switch of lateralization, and the remaining four patients had a regionalized ictal EEG. Out of the 49 patients with unilateral mesial temporal MRI pathology, two had independent left and right ictal EEG in different seizures, 11 had bitemporal asynchronous ictal EEG or switch of lateralization, nine patients had a non-lateralized ictal EEG, and 27 patients had a regionalized ictal EEG. By visual analysis, one of the eight patients without lateralized mesial temporal MRI pathology had bilateral mesial temporal lobe atrophy in MRI; this patient, with a history of encephalitis, had bitemporal IED and a left temporal ictal EEG. No significant correlations could be shown between recollection of an aura and duration of
Fig. 2. Interictal epileptiform discharges (IED) as a function of postictal memory of an aura in EEG/video monitoring.
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
207
Fig. 3. Unilateral mesial temporal pathology in MRI as a function of postictal memory of an aura in EEG/video monitoring.
epilepsy, prose memory performance, and age of occurrence of possible etiological factor (Mann – Whitney test, P [ 0.328). Aura memory in video/EEG monitoring was also a function of the reliability of the localization and lateralization of the mesial TLE. Failure to recall an aura after each seizure occurred significantly more often in patients with unreliably diagnosed lateralization of mesial TLE (54.5%, six of 11 patients) than in patients with reliably diagnosed lateralization of mesial TLE (right TLE, 4.8% (one of 21 patients); left TLE, 15.4% (four of 26 patients); Fisher’s exact test, two-sided, P= 0.031). Significant differences between right and left TLE could not be found with regard to aura memory. The diagnosis of temporal lobe epilepsy as opposed to extratemporal epilepsy could be confirmed in all eight patients without lateralized MRI pathology: four of the eight patients had significantly lateralized temporal hypometabolism, and three had non-lateralized temporal hypometabolism; the remaining patient had been studied with subdural and depth electrodes in another epilepsy center, with the result of independent right and left temporal subclinical and clinical EEG seizure patterns.
4. Discussion This study shows that patients with mesial TLE, without or with only rare auras (group 3 patients), had signs of bilateral mesial temporal lobe dysfunction. Evidence is given by the following findings: (1) interictal scalp EEG (bitemporal IED); (2) ictal scalp EEG (bilateral independent, asynchronous EEG seizure patterns or switch of lateralization); and (3) the absence of lateralized MRI pathology of the mesial temporal lobes. The correlation of group 3 patients with bitemporal IED is consistent with Steinhoff’s findings of a significantly rarer aura in patients with bitemporal versus unitemporal IED in scalp EEG (Steinhoff et al., 1995). With regard to the ictal EEG, possible reasons for not having auras or not recalling auras could be as follows. (1) Rapid propagation of the epileptiform discharge to the contralateral temporal lobe with bilateral inactivation of the mesial temporal structures that are essential for memory encoding. This could be related to epileptogenicity of the contralateral temporal lobe or relative severity of the seizures. Both mechanisms would facilitate bilateral hippocampal seizure spread. (2)
208
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
Epileptic discharge, inactivating one mesial temporal structure with morphological damage to the contralateral hippocampus (this would also lead to functional inactivation of both mesial temporal structures). (3) Rapid propagation of the epileptic discharge into both temporal lobes from an extratemporal epileptogenic zone, e.g. the parietal or occipital lobes. Our findings of bilateral, independent, asynchronous EEG seizure patterns or switch of lateralization of the EEG seizure pattern in surface EEG in group 3 patients corroborate a previous study of a heterogeneous group of patients with focal epilepsy (Schulz et al., 1995). In this study, we found a correlation of ictal scalp EEG propagation to the contralateral hemisphere with amnesia of the previously documented aura, but we had addressed individual seizures and had considered the ictal EEG only during the aura. Furthermore, we did not differentiate asynchronous bitemporal ictal EEG patterns and switch of lateralization from EEG seizure patterns with identical frequency over both hemispheres so that a bilateral EEG seizure pattern during the aura could have resulted from volume conduction to the contralateral anterior and mesial (sphenoidal) electrodes or from a diffuse EEG seizure pattern with later regionalization according to the principles stated by Risinger et al. (1989). The present study defines the EEG patterns of switch of lateralization and bilateral asynchrony, both of which can be regarded as an independent EEG seizure pattern involving the contralateral temporal lobe in patients with mesial TLE (Schulz et al., 2000). Severity of the seizure was not a likely cause for failure to recollect an aura in group 3 patients because secondarily generalized tonic – clonic seizures were also rare in this patient group. Actually, the only patient in this group who had secondary generalization also had four additional psychomotor seizures without recollection of an aura. On the other hand, a switch of lateralization of the EEG seizure and an asynchronous EEG seizure pattern over both temporal lobes were found significantly more often in patients without aura, consistent with our hypothesis of a bitemporal dysfunction.
A discrete lesion of the contralateral temporal lobe or secondary contralateral epileptogenesis are possible causes of bitemporal dysfunction. Morrell (1989) established the concept of secondary epileptogenesis by stimulation studies in animals and by studies of epileptic patients with unilateral tumors. In this concept, frequent ictal propagation to the contralateral temporal lobe is thought to cause contralateral IED. Bilateral independent EEG seizure patterns could also result from a more extensive damage to the ipsilateral mesial temporal lobe, which might facilitate faster contralateral propagation. In this case, patients with marked mesial temporal sclerosis on MRI should have bitemporal ictal patterns more often than patients without or with only moderate unilateral mesial pathology. However, we found no correlation of independent bitemporal ictal propagation with unilateral MRI pathology. Therefore, independent epileptogenicity of the contralateral temporal lobe or a different mode of propagation across the frontal lobes (Lieb et al., 1991) or the posterior hippocampal commissure (Gloor et al., 1993) might explain the absence of auras in this group. Propagation from an extratemporal epileptogenic zone into both temporal lobes is difficult to exclude in non-lesional epilepsies. Temporal hypometabolism in PET in seven of eight patients without unilateral mesial temporal MRI pathology and investigation with subdural and depth electrodes in the remaining patient make an extratemporal epileptogenic zone less probable. Bitemporal dysfunction can result in poor memory functions in general. Vargha-Khadem et al. (1997) described severe deficits of episodic memory in three patients with marked bitemporal MRI pathology. In our study, aura history and aura recall were not significantly related with the recall of prose passages as an indicator of episodic memory. Group 3 patients significantly more often had bitemporal IED and bitemporal independent ictal EEG; but they also significantly more often had no unilateral mesial temporal MRI pathology. With the exception of one patient, we did not find marked bitemporal pathology in this group, which would have been detectable by visual analysis similar to the patients of Vargha-
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210
Khadem et al. We suppose that more discrete bitemporal pathology that is not detectable by visual analysis causes no severe interictal deficits of episodic memory, but only the comparatively discrete signs of amnesia of an aura or failure to experience an aura. According to our alternative hypothesis, the interictal absence of severe deficits of episodic memory in Group 3 patients could also be explained as an exclusively ictal phenomenon by rapid propagation and inactivation of the contralateral mesial temporal lobe. MRI volumetry of both hippocampi, with T2 relaxometry, and with MR spectroscopy could be of interest to assess discrete bitemporal mesial pathology. It is also theoretically possible that patients without aura during monitoring still suffer from transient postictal memory deficits. Bergin et al. (1995) found no memory deficits during monitoring after 48 h in spite of complex partial seizures, secondarily generalized tonic – clonic seizures or both in the interval. Helmstaedter et al. (1994) described verbal and non-verbal learning deficits with a duration of 30 min to 1 h after complete reorientation. Helmstaedter et al. tested postictal learning but not postictal retrieval of preictally encoded episodic material. We think that reorientation with respect to person, location and time represents intact episodic retrieval so that the failure to recall an aura postictally should reliably indicate the absence of an aura or the failure to encode an aura at the beginning of the seizure. All our patients were asked for memory of an aura after renewed postictal reactivity, reorientation and resolved possible postictal aphasia. Postictal memory deficits in some patients because of insufficient encoding could indicate bitemporal mesial dysfunction according to our hypothesis. The relatively poor surgical outcome in patients with switch of lateralization and asynchronous EEG seizures (Schulz et al., 2000) also supports the hypothesis of bitemporal epileptogenicity in group 3 patients. Another explanation for a worse prognosis could be a more posterior pathology in these patients together with an incomplete resection of the most posterior parts of the hippocampus and parahippocampus, so that propagation from remaining epileptogenic tissue across the
209
posterior hippocampal commissure might occur. On the other hand, postoperative remains of epileptogenic tissue are more probable in patients with mesial temporal sclerosis visible on MRI than in patients without visible unilateral mesial pathology. Our study shows that patients with mesial TLE without or with only rare auras in their history, who had no auras during non-invasive EEG/video monitoring, significantly more often had bitemporal IED, ictal patterns suggestive of a bitemporal independent epileptogenicity, absence of lateralized mesial temporal MRI pathology, and a worse outcome, which together indicate a bilateral mesial temporal disease or a different seizure propagation mode. Patients with the aforementioned clinical features might therefore form a subgroup of patients with mesial TLE who often need invasive monitoring with subdural strips or depth electrodes.
Acknowledgements The authors thank Ralf Dernbach, Martina Fru¨nd, and their teams of EEG technicians and nurses for their commitment in the ictal and postictal testing of the patients.
References Bergin, P.S., Thompson, P.J., Fish, D.R., Shorvon, S.D., 1995. The effect of seizures on memory for recently learned material. Neurology 45, 236– 240. Cyr, L., Francis, K., 1992. Measures of clinical agreement for nominal and categorical data. The kappa coefficient. Comput. Biol. Med. 22, 239– 246. Ebner, A., Hoppe, M., 1995. Noninvasive electroencephalography and mesial temporal sclerosis. J. Clin. Neurophysiol. 12, 23 – 31. Engel, J., 1994. Bilateral temporal lobe epilepsy. In: Wolf, P. (Ed.), Epileptic Seizures and Syndromes. John Libbey, London, pp. 359– 368. Gloor, P., Salanova, V., Olivier, A., Quesney, L.F., 1993. The dorsal hippocampal commissure. An anatomically identifiable and functional pathway. Brain 116, 1249– 1273. Halgren, E., Wilson, C.L., Stapleton, J.M., 1985. Human medial temporallobe stimulation disrupts both formation and retrieval of recent memories. Brain Cogn. 4, 287– 295.
210
R. Schulz et al. / Epilepsy Research 43 (2001) 201–210 Schulz, R., Lu¨ders, H.O., Noachtar, S., May, T., Sakamoto, A., Holhausen, H., Wolf, P., 1995. Amnesia of the epileptic aura. Neurology 45, 231– 235. Schulz, R., Lu¨ders, H.O., Hoppe, M., Tuxhorn, I., May, T., Ebner, A., 2000. Interictal EEG and ictal scalp EEG propagation are highly predictive of surgical outcome in mesial temporal lobe epilepsy. Epilepsia 41, 564– 570. Scoville, W.B., Milner, B., 1957. Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatr. 20, 11 – 21. Sperling, M.R., Lieb, J.P., Engel, J., Crandall, P.H., 1989. Prognostic significance of independent auras in temporal lobe seizures. Epilepsia 30, 322– 331. Squire, L.R., 1986. Mechanisms of memory. Science 232, 1612– 1619. Steinhoff, B.J., So, N.K., Lim, S., Lu¨ders, H.O., 1995. Ictal scalp EEG in temporal lobe epilepsy with unitemporal versus bitemporal interictal epileptiform discharges. Neurology 45, 889– 896. Vargha-Khadem, F., Gadian, D.G., Watkins, K.E., Connelly, A., Van Paesschen, W., Mishkin, M., 1997. Differential effects of early hippocampal pathology on episodic and semantic memory. Science 277, 376– 380.
Helmstaedter, C., Elger, C.E., Lendt, M., 1994. Postictal courses of cognitive deficits in focal epilepsies. Epilepsia 35, 1073– 1078. Lieb, J.P., Dasheiff, R.M., Engel, J., 1991. Role of the frontal lobes in the propagation of mesial temporal lobe seizures. Epilepsia 32, 822– 837. Mathern, G.W., Babb, T.L., Vickrey, B.G., Melendez, M., Pretorius, J.K., 1995. The clinical–pathogenetic mechanisms of hippocampal neuron loss and surgical outcomes in temporal lobe epilepsy. Brain 118, 105–118. Morrell, F., 1989. Varieties of human secondary epileptogenesis. J. Clin. Neurophysiol. 6, 227–275. Palmini, A.L., Gloor, P., Jones-Gotman, M., 1992. Pure amnestic seizures in temporal lobe epilepsy. Brain 115, 749– 769. Penfield, W., Jasper, H., 1954. Epilepsy and the Functional Anatomy of the Human Brain. Little, Brown & Co, Boston, MA, p. 528. Risinger, M.W., Engel, J., Van Ness, P.C., Henry, T.R., Crandall, P.H., 1989. Ictal localization of temporal lobe seizures with scalp/ sphenoidal recordings. Neurology 39, 1288– 1293.
.