Epilep\i[i, 40(9):1261-1268, 1999

Lippincott Williams & Wilkim, Inc., Philadelphia

0 International League Again\t Epilepsy

Clinical Research

Interictal Epileptiform Discharges in Temporal Lobe Epilepsy

Due to Hippocampal Sclerosis Versus Medial Temporal
Lobe Tumors

H. M. Hamer, I. Najm, A. Mohamed, and E. Wyllie

The Cleveland Clinic Foundation, Section of Epilepsy and Sleep Disorders, Department of Neurology, Cleveland, Ohio, U.S.A.

Summary: Purpose: It remains controversial whether a spe- dian, 72.1 %; p < 0.001 ; sphenoidal electrode alone, 24.8%; p <
cific pattern of interictal epileptiform activity exists that may 0.001). The HS group had significantly fewer extratemporal
help to differentiate temporal lobe epilepsy (TLE) due to hip- spikedsharp waves (median, O.O), as compared with the tumor
pocampal sclerosis (HS) from other forms of TLE. In this group (10.0%: p < 0,001). At least 90% of the interictal dis-
study, we characterized the distribution of interictal epilepti- charges were located in the anterior temporal region in 20
form discharges in TLE due to HS as compared with those in (95.2%) of 21 HS patients, but in none of the tumor patients (p
patients with tumors restricted to the medial temporal lobe < 0.001). Bilateral temporal discharges were found in nine
structures. (42.9%) of 21 patients with HS and in two (22.2%) of nine
Methods: The study included 21 adult patients with unilateral tumor patients (p = 0.42).
HS who remained seizure free (> 1 year) after anterior temporal Conclusions: We conclude that ipsilateral interictal epilepti-
lobectomy with amygdalohippocampectomy. Patients with form discharges outside the anterior temporal region are rare
“dual pathology” were excluded. The comparison group con- (<lo%) in adults with intractable TLE due to unilateral HS.
sisted of nine patients with tumors restricted to the amygdala Frequent posterior or extratemporal sharp waves may detract
and hippocampus. All patients underwent video-EEG monitor- from the certainty of this diagnosis in complicated cases. These
ing preoperatively by using 39 scalp electrodes (including the restricted epileptiform discharges suggest a smaller irritative
10-10 system over both temporal regions) and bilateral sphe- zone in HS as compared with medial tumors, or a more orga-
noidal electrodes. nized activity associated with intrinsic hippocampal disease.
Results: The HS patient group showed a significantly higher Bilateral epileptiform discharges were not uncommon in both
percentage of ipsilateral epileptiform discharges maximal at groups. Key Words: Temporal lobe epilepsy-Hippocampal
anterior temporal electrodes (median, 97.0%; sphenoidal elec- sclerosis-Tumor-Interictal epileptiform discharges-EEG.
trode alone, 88.1 %), as compared with the tumor group (me-

Temporal lobe epilepsy (TLE) is one of the most com- prognosis for becoming seizure free with antiepileptic
mon forms of localization-related epilepsy (1-3) and in drugs (AEDs) (9). In these refractory patients, anterior
the majority of TLE, the medial temporal lobe structures temporal lobectomy with removal of the medial struc-
including amygdala and hippocampus constitute the epi- tures offers an effective treatment, rendering -70% of
leptogenic region (4,5). By using volumetric measure- these patients seizure free and 20% clinically improved
ments, magnetic resonance imaging (MRI) findings have (10). These findings have led to a syndromatic definition
been shown to be highly accurate for in vivo diagnosis of of MTLE due to HS (9). The features of this syndrome
hippocampal sclerosis (HS) as the underlying pathology include characteristic seizure semiology (epigastric or
(6-8). A relatively large subset of patients with medial psychic aura followed by clouding of consciousness, oro-
temporal lobe epilepsy (MTLE) due to HS have a poor alimentary and manual automatisms, and autonomic phe-
nomena) (1 1,12) and characteristic electroencephalogra-
phy (EEG) findings (regional temporal or lateralized ic-
Accepted February 9, 1999. tal EEG recordings and uni- or bilateral temporal
Address correspondence and reprint requests to Dr. I. Najm at The slowing and epileptiform discharges on interictal EEG)
Cleveland Clinic Foundation, Gepartment of Neurology, S5 1, 9500 (13-15).
Euclid Avenue, Cleveland, OH 44195, U.S.A. E-mail: najmi@ccf.org
Presented in part at the 50th Annual Meeting of the American Acad- Although temporal spikes and sharp waves are the
emy of Neurology, Minneapolis, April 25-May 2, 1998. hallmark of the interictal EEG of MTLE, it is controver-

1261
1262 H. M. HAMER ET AL.

sial whether a specific pattern of interictal activity exists

that may help to confirm MTLE due to HS in challenging
cases (15-17). In th.is study, we aimed to characterize
further the pattern of interictal epileptiform discharges in
MTLE due to HS. We studied the preoperative interictal
discharges of patients with HS who became seizure free
after limited anterior temporal resection with removal of
the medial temporal lobe structures. This was then com-
pared with the interictal discharges of a highly selected
group of patients with tumors restricted to the medial
temporal lobe structures (amygdala and/or hippocam-
pus). Thus we compared two pathologic conditions that
affect the same anatomic region with potentially distinct
mechanisms of epileptogenicity. Differences in the pat-
tern of epileptiform discharges may help in the presur-
gical evaluation to confirm or detract from a certain di-
agnosis, such as MTLE due to HS.
FIG. 1. Left lateral view of scalp electrode montage according to
the International 10-10 system over the temporal lobe area and
PATIENTS AND METHODS according to the 10-20 system over the rest of the scalp. Montage
on right side included corresponding electrodes with even nurn-
During the period of 1992 to 1996, a total of 186 bers. Additionally, bilateral sphenoidal electrodes were placed in
patients with medically intractable TLE underwent ante- all patients (left, Sp,; right, S,,J.
rior temporal lobe resections at the Cleveland Clinic
Foundation. The prolonged scalp video-EEG recordings
The first author (H.M.H.), who reviewed all the pa-
from two subgroups of these patients were collected ret-
tients’ EEGs, was blinded to the pathologic diagnosis
rospectively and reanalyzed. One group consisted of pa-
during the EEG analyses. If >50 discharges were avail-
tients with TLE due to HS, and the other group included
able, the epileptiform activity was sampled evenly
patients with medial temporal lobe tumors. Additional
throughout the entire length of monitoring. A spike or
data were obtained through chart reviews and telephone
sharp wave had duration of <200 ms and could be dis-
interviews with the patients.
tinguished by its structure and amplitude from back-
Presurgical evaluation ground activity (Fig. 2) (19). A slow wave typically fol-
Preoperative evaluation of all patients included com- lowed. Nonepileptogenic epileptiform activity, such as
plete medical history and neurologic examination, con- benign epileptiform transients of sleep, was excluded
tinuous scalp video-EEG monitoring, and MRI studies. (20). Interictal epileptiform discharges (spikes or sharp
The EEG was digitally recorded by using 39 scalp waves) were visually selected ( 4 0 spikeskharp Gaves
electrodes and additional left and right sphenoidal elec- ipsilateral to the lesion per patient). The upstroke of the
trodes (Spl and Sp2)in all patients. The scalp electrodes spikelsharp wave was analyzed with respect to location
were placed according to the International 10-10 system of the maximal negativity and amplitude and field dis-
over both temporal areas and according to the 10-20 tribution of the discharge (21). For mapping of the nega-
system over the rest of the scalp (Fig. 1) (18). Electrode tivity, a referential montage was used with the contralat-
impedances were kept <5 kOhm. During acquisition, the era1 occipital electrode (0, or 0,) as reference. Bitem-
bandpass was 0.5-70 Hz, and sensitivity was 15 kV/mm. poral synchronous discharges were lateralized to the side
The sampling rate was at 200 Hz. with the higher voltage. Postictal recordings were ex-
In this study, only interictal epileptiform discharges cluded from evaluation until the EEG returned to base-
ipsilateral to the lesion were considered for the evalua- line. Because of the limited number of interictal spikes
tion of discharge distributions. For the purpose of sim- and sharp waves, we did not control for state of con-
plifying the reporting of study results, homologous right sciousness (awake/sleep) while evaluating the dis-
(e.g., F,) and left electrodes (e.g., F,) were combined in charges.
one expression (e.g., F,,,) because the symptomatic side MRI studies were performed 4-6 weeks before sur-
could be either the left or the right side in each patient. In gery on all patients, by using a Siemens 1.5-T Vision
addition, electrodes S,,,,, F,,,, F,,,,, FT,,,, and FT,,,o system. Multiple coronal T I and T, images, sagittal T I
were referred to as anterior temporal electrodes, elec- images, and axial T,, spin echo and FLAIR images were
trodes T,,,, T9,,,, TP,,,, TP,,,,, and P,,, were referred to acquired on all patients. Image processing was per-
as lateral temporal electrodes, and the rest of scalp elec- formed by using ISG Allegro software (ISG Technolo-
trodes were defined as extratemporal electrodes (Fig. 1). gies Inc., Toronto, Ontario, Canada) running on a Silicon
 EPILEPTIFORM DISCHARGES IN MTLE 1263

B
Fpl-01
I-v--J--
100 pv
F7-01
SP1-01
T7-01
P7-01

FIG. 2. lnterictal epileptiform dis- Fp2-01

charges. A: Sharp wave maximal
at the right sphenoidal electrode F8-01
&). B: Sharp wave maximal at
the left lateral temporal electrode sp2-01
T,. C: Sharp wave maximal at the
left extratemporal electrode F,. T8-01
Electrode montage: Referential
montage using left occipital elec- P8-01
trode 0, as reference. Montage
includes left and right sphenoidal FP1-01
electrodes (S,,, SP2) as well as
T,, and T,, of the International F3-01
10-10 system.
C3-01
P3-01
FP2-01
F4-01
C4-01
P4-01

Graphics workstation. For each subject, right and left system over both temporal lobes and additional bilateral
hippocampal formation volumes were measured from the sphenoidal electrodes (Fig. 1); (e) interictal spikedsharp
coronal sections by following the anatomic guidelines waves recorded; and (f) follow-up of 2 1 year after sur-
previously outlined by Watson et al. (22). To define hip- gical resection (range, 1 4 . 5 years).
pocampal atrophy, normative MRI data were obtained All HS patients underwent prolonged video-EEG
from a sample of 24 healthy adult volunteers. Based on monitoring for 4.8 f 2.1 days (range, 1-8 days), and a
these acquired normative data, right (R) and left (L) hip- total of 582 interictal epileptiform discharges was evalu-
pocampal atrophy were defined as R L volume ratios ated. HS was confirmed by postoperative histopathologic
<0.89 and >1.13, respectively. Gadolinium was admin- analysis of surgical specimens in nine patients. Failure to
istered to patients with identifiable or suspected tumor identify the histopathology in the remaining patients may
pathology. have been due to sampling error, as few of the early
resections were performed en bloc. Patients with addi-
Hippocampal sclerosis group
tional evidence of cortical dysplasia or foreign-tissue le-
In the HS group, 21 patients with partial seizures (age
sions on MRI or histopathology were excluded from this
33.4 -+ 10.6years; Table 1 ) were selected from a group of
group.
145 consecutive patients who underwent temporal lobec-
tomy because of medically intractable TLE of nontumor Tumor group
origin according to the following criteria: (a) unilateral The tumor group consisted of nine highly selected
hippocampal atrophy by MRI volumetric study; (b) ab- age-matched patients with medically intractable partial
sence of abnormalities other than unilateral HS on MRI seizures (Table 1) due to tumors involving only the me-
or pathology; (c) seizure-free outcome after limited an- dial temporal lobe structures as detected by MRI and con-
terior temporal resection with removal of the medial f m e d by postoperative histopathologic analyses of surgi-
structures; (d) prolonged video-EEG monitoring with cal specimens. The patients were selected from a group of
electrode montage according to the International 10-10 41 consecutive patients who underwent temporal lo-

Epilepsia, Vol. 40, No. 9, 1999

1264 H. M. HAMER ET AL.

TABLE 1. Distribution of ipsilateral epileptiform discharges in patients with hippocampal sclerosis versus medial
temporal tumors
Ant. Lat. Postoperative follow-up
Age Side of Bilateral Number of Temp. Temp. Extratemp.
No. (yr) Sex disease Frequency" discharges spikes/s.w. Sp112 (%) (%I Outcome Duration
Hippocampal sclerosis group
I 21 F L 4 20 25 100 0 0 Sz. free 21 mo
2 31 M L 0.07 11 72.7 90.9 9.1 0 Sz. free 30 mo
3 3 6 M R 1 - 27 44.4 100 0 0 Sz. free 50 mo
4 31 F L 0.03 - 19 89.5 100 0 0 Sz. free 30 ma
5 4 4 M R Rare Yes 4" 100 100 0 0 Sz. free 13 mo
6 16 M L 1 - 20 65 90 10 0 Sz. free 54 mo
7 27 F R 1.5 Yes 32 90.6 96.9 3.1 0 Sz. free 42 mo
8 25 F L 30 Yes 50 90 I00 0 0 Sz. free 34 mo
9 21 F L 0.03 - 38 84.2 92.1 7.9 0 Sz. free 51 mo
10 44 M R 1.5 Yes 50 80 90 8 2 Sz. free 21 ma
11 46 M R 0.75 - 15 80 100 0 0 Sz. free 55 mo
12 42 F L 0.05 Yes 30 86.7 96.7 3.3 0 Sz. free 36 mo
13 36 M L Rare 11 63.6 90.9 0 9.1 Sz. free 24 mo
14 39 F R 5 - 33 97 97 0 3 Sz. free 27 mo
15 49 F L 10 Yes 50 94 100 0 0 Sz. free 12 mo
16 26 F R 0.5 Yes 30 96.7 96.7 3.3 0 Sz. free 12 ma
17 54 F L 0.017 - 11 90.9 90.9 9.1 0 Sz. free 26 mo
18 38 M R 0.5 Yes 18 94.4 100 0 0 Sz. free 37 mo
19 24 F R 0.067 - 20 95 100 0 0 Sz. free 50 mo
20 31 M L 2 Yes 50 96 100 0 0 Sz. free 18 mo
21 21 F L 0.1 - 43 81.4 86. I 2.3 11.6 Sz. free 31 mo
Tumor group
I 38 M L Rare Ib 0 0 100 0 Sz. free' 24 mo
2 5 3 M R 0.2 - 20 20 90 0 10 N/A N/A
3 39 F R 1 Yes 14 7.1 57.1 42.8 0 Sz. free 12 mo
4 3 7 M R 12 Yes 50 60 82 8 10 50% impr.d 30 mo
5 3 6 M R Rare - 21 61.9 71.4 23.8 4.8 Sz. free 20 mo
6 36 F L I .5 27 29.6 48.2 44.4 7.4 Sz. free 50 mo
7 4 7 M L 0.033 21 0 66.7 14.3 19.1 Sz. free 28 mo
8 22 F L Rare I1 63.6 72.7 9.1 18.2 Sz. free 12 mo
9 15 F L 10 12 8.3 75 8.3 16.7 80% impr." 25 mo

s.w., sharp waves; SP1/2, sphenoidal electrodes; ant. temp., anterior temporal electrodes including sphenoidal electrode; lat. temp., lateral temporal
electrodes; extratemp., extratemporal electrodes; F, female; M, male; L, left; R, right; sz. free, seizure free; N/A, not available; impr., improved.
'' Frequency given as spike/sharp wave per inin (as reported by the reviewing neurophysiologist).
I' Dtie to insufficient number of spikedsharp waves, patient was excluded from statistical comparison of spikelsharp wave distributions between
both groups.
One isolated seizure on accidental abrupt drug discontinuation.
'' Incomplete tumor resection on postoperative MRI.

bectomy for medically intractable TLE due to temporal continuation of carbamazepine (CBZ). One patient was
lobe tumors. In seven patients, the tumor was confined to lost to follow-up. In the remaining two patients, there
the region of amygdala and hippocampus only. In the was 5040% decrease in seizure frequency, and the post-
remaining two patients (patients 6 and 9, Table l), the operative MRI showed evidence for an incomplete tumor
tumor involved also the parahippocampal gyrus to a very resection. Prolonged video-EEG monitoring was per-
limited extent. Tumor pathology revealed low-grade as- formed for 6.9 k 2.6 days (range, 4-13 days), and 177
trocytoma (patients 4 and 7; 22.2%), low-grade glioma interictal epileptiform discharges were evaluated.
(patients 3 and 5 ; 22.2%), oligodendroglioma (patient 2;
ll.l%), hamartoma (patients 1, 8, and 9; 33.3%), and Statistical analysis
ganglioglioma (patient 6; 1 1.1%). Patients with tumors Because most of the variables in the study were not
extending outside the medial temporal lobe structures as normally distributed (Shapiro-Wilk's W test), the non-
detected by MRI or histopathology were excluded from parametric Mann-Whitney U test, and Fisher's Exact test
the study. None of the patients had dual pathology with were used for group comparisons. In addition, confi-
both tumor and HS on MRI or histopathologic analysis. dence intervals were calculated for each of the median
Only patient 6 had microscopic cortical dysplasia asso- values by using a technique described by Cambell and
ciated with a ganglioglioma. Six patients became seizure Gardner (23). Confidence intervals for a quantile such as
free postoperatively including one patient who experi- the median do not necessarily allow calculation of an
enced one isolated seizure after unplanned, abrupt dis- interval with exactly 95% confidence. As a result, the

Eplepscu, Vol. 40, No. 9, 1999

 EPILEPTIFORM DISCHARGES IN MTLE 1265

confidence intervals reported in Table 1 are at the 96% Furthermore, 19 (90.5%) HS patients had >60% of their
level for the HS group and the 93% level for the tumor discharges maximal at the sphenoidal electrodes com-
group. Only patients with 2 10 recorded spikes or sharp pared with only two (22.2%) tumor patients (p < 0.001).
waves ipsilateral to the lesion were included in statistical Nine (42.9%) of the 21 HS patients had infrequent lateral
group comparisons with respect to epileptiform dis- temporal epileptiform discharges (range, 2.3-lo%), and
charge distributions. Therefore, patient 5 of the HS group only four (19.1 %) HS patients showed extratemporal
and patient 1 of the tumor group were excluded from spikes or sharp waves (range, 2-11.6%; Table 1). In
these comparisons (Table 1). contrast, seven (77.8%) of the nine tumor patients had
extratemporal discharges (range, 4.8-19.1%; p = 0.002
RESULTS compared with the HS group; Table 1).
In the HS group, extratemporal spikekharp wave
Bilateral independent temporal discharges were found
maximums involved electrodes AF,,,, and in one case
in nine HS patients (42.9%; Table 1). In seven (77.8%)
(patient 21), electrode F, (Fig. 3). The extratemporal
of nine patients, the discharges were predominantly on
discharges of the tumor group were more widespread,
the side of the disease, with ratios ranging from 2:1 to
involving frontal and central leads (F,,,,, F,,, AF,,,,
1OO:l. In two (22.2%) patients, spikes and sharp waves
FC,, and C5).
were more frequent contralateral to the side of the HS at
There were no significant differences in frequency
ratios of 1:3 and 2:3. Two (22.2%) of the nine tumor
(Table 2) and amplitude of the interictal discharges be-
patients (p = 0.42 compared with HS group) showed
tween patients with HS and medial temporal lobe tumors
bilateral independent discharges that were predominantly
(Tu) (median of amplitude: HS, 80.3 FV; Tu, 73.3 FV;
ipsilateral to the side of the medial temporal lobe tumor
p = 0.48). Discharge frequency did not correlate with
at ratios of 3: 1 and 100:1. In patients with bilateral dis-
age, gender, duration of epilepsy, or side of disease in
charges, spikes and sharp waves were restricted to ho-
both groups (Table 1).
mologous electrodes. There was no significant difference
in the duration of the epilepsy between patients with
bilateral (median, 18.5 years) versus unilateral dis- DISCUSSION
charges (median, 16.0 years; p = 0.35).
The HS group showed significantly more discharges Our results showed that interictal epileptiform dis-
maximal at the ipsilateral sphenoidal electrode than did charges in patients with HS are highly restricted to an-
the tumor group (Table 2). Moreover, patients with HS terior temporal electrodes. Moreover, the data suggested
had a median of 97.0% of the discharges located in the a disease-specific distribution of epileptiform discharges
anterior temporal region (S,,,,, F,,,, Fsrl0,FT,,,, FT,,,,), in HS. This pattern contrasts with epileptiform dis-
as compared with only 72.1% in the tumor group (p < charges in patients with tumors restricted to the amyg-
0.001; Table 2). The HS group showed significantly dalahippocampal area who had disease in the same re-
fewer extratemporal spikeshharp waves, as compared gion but showed a more widespread distribution of in-
with the tumor group. terictal activity. Patients with dual pathology were
In 20 (95.2%) of the 21 HS patients, 290% of the excluded, as our study compared the pattern(s) of epi-
interictal epileptiform discharges were located in the an- leptogenicity of two distinct pathologic conditions in-
terior temporal region (Table 1). None of the patients volving the medial temporal lobe and associated with
with medial temporal lobe tumors had >90% of their medically intractable TLE. As a rule, patients with HS in
interictal spikekharp waves in this area (range, 0-90%; this series had >90% of their interictal epileptiform dis-
Fisher's Exact test: p < 0.001 compared with HS group). charges in the anterior temporal region. Therefore fre-

TABLE 2. Characteristics of interictal epileptijorm discharges in patients with

hippocampal sclerosis versus medial temporal lobe tumors
Median (CI)"
HS Tu
Spikeslsharp waves (n = 20) (n = 8) p Value
Maximum at sphenoidal electrode (%) 88.1 (80.0-94.0) 24.8 (7.1-61.9) <0.001
Maximum at anterior temporal electrodes (a) 97.0 (90.9-100.0) 72.1 (57.1-82.0) <0.001
Maximum at lateral temporal electrodes (%) 0.0 (0.0-3.3) 11.7 (8.0-42.8) 0.003
Maximum at extratemporal electrodes (%) 0.0 (0.0-0.0) 10.0 (4.8-18.2) <0.001
Frequency (per min) 0.6 (0.1-1.5) 0.6 (0.0-10.0) 0.92

Anterior temporal electrodes: S,,,,, F7/8,F,/,", FT7/*,F,,,,,.

HS, Hipposcampal sclerosis group; Tu, Mesial temporal tumor group.
' Confidence interval (CI) was 96% for HS group and 93% for Tu group.

Epiiepsia, Vol. 4Q No. 9, I999

I266 H. M . HAMER ET AL.

1
v) 80
a
En
$ 70
c
60

-
5
Q
U
50

40 -~ _ _ - FIG. 3. Distribution of ipsilateral in-

terictal epileptiform discharges in hrp-
.
. ~ pocampal sclerosis (HS; n = 582) ver-
sus medial temporal lobe tumors (Tu,
~ ~~ - -- n = 177).

A
---
Anterior temporal Lateral temporal Extra-temporal electrodes

quent posterior or extratemporal sharp waves may de- the two patients who continued to exhibit seizures after
tract from the certainty of the diagnosis of HS. surgery.
These restricted anterior temporal discharges in our In extracranial EEG recordings, interictal discharges
HS patients confirmed previous studies (1 5,16) and may may have a more reliable lateralizing value than ictal
suggest a smaller irritative zone (5,24) as compared with changes in TLE (31), and they almost always predict
medial temporal lobe tumors. Alternatively, there may be seizure origin if they are exclusive or clearly preponder-
more organized interictal activity associated with intrin- ant in a single region (32). Similarly, about half of the HS
sic hippocampal disease. These data are in agreement patients in our study had epileptiform discharges only on
with previous hypotheses on the generators of interictal the side of the disease. In the other half, the discharges
epileptiform activity in HS (25,26). By using three- were bilateral but usually with a predominance to the
dimensional multiple dipole modeling, Baumgartner et side of HS. It remains unclear why bilateral discharges
al. (26) identified two possible sources of interictal occur in hippocampal sclerosis. This may represent ad-
spikes in HS patients. The first source involved the me- ditional less severe contralateral HS (33) and indepen-
siobasal aspect of the temporal lobe (hippocampus and dent bitemporal excitability (34), or secondary epilepto-
parahippocampal gyrus) and was followed within 40 ms genesis reflecting a “mirror focus” (3 1). The presence of
by activation of the anterior temporal lobe neocortex. bilateral independent discharges in tumor patients who
The rare occurrence of lateral temporal and fiontal became seizure free after tumor resection supports the
spikeshharp waves in HS as recorded by surface elec- hypothesis of mirror foci in an “intermediate state” of
trodes may be due to spread of epileptic discharges to the secondary epileptogenesis (35) rather than independent
temporal neocortex or to the orbitofrontal cortex via the bilateral disease.
limbic system. Previous studies revealed favorable seizure-free out-
On the other hand, in this study, patients with tumors come despite bitemporal independent spikes or sharp
restricted to the medial temporal lobe structures showed waves when MRI-based hippocampal atrophy was seen
a more widespread interictal activation of the lateral and on the side of EEG seizure onset (15,36). These findings
posterior temporal and extratemporal cortex. Their irri- were independent of the duration of epilepsy. Our results
tative zone of interictal epileptiform discharges is con- are in concordance with these reports, as nine (42.8%) of
tiguous with, but extends beyond the structural lesion. our patients who became seizure free after epilepsy sur-
This may be caused by local neuronal injury, edema, gery ipsilateral to the hippocampal atrophy had bitem-
ischemia, or other electrical andor biochemical effects poral interictal discharges. This included two patients
of the structural lesion on susceptible neighboring brain with the preponderance of the interictal discharges con-
tissue (27,28). Recent studies have shown that complete tralateral to the HS. Therefore patients with bitemporal
tumor removal is highly associated with seizure freedom, discharges and unilateral hippocampal atrophy may ben-
whereas the likelihood of complete remission is small in efit from temporal lobectomy (37,38), although the ma-
subtotal resections (29,30). Accordingly, in our series, jority of postoperatively seizure-free patients showed
postoperative MRI showed subtotal tumor resection in discharges preoperatively limited to only one temporal
 EPILEPTIFORM DISCHARGES IN MTLE 1267

lobe (39). The presence of hippocampal atrophy detected REFERENCES

by MRI ipsilateral to the side of seizure origin may serve
I . Gastaut H, Gastaut JL, Gontalves e Silva GE, Fernandez Sanchez
as an important independent predictor of favorable out- GR. Relative frequency of different types of epilepsy: a study
come in these patients. employing the classification of the International League Against
Including only surgical cases allowed us to define a Epilepsy. Epilepsia 1975;16:457-61.
2. Loiseau P, Duch6 B, Loiseau J. Classification of epilepsies and
strict HS group based on seizure-free outcome after hip- epileptic syndromes in two different samples of patients. Epilepsiu
pocampal resection in addition to MRI volumetric evi- 1991;32:303-9.
dence of hippocampal atrophy. This helped to compen- 3. Manford M, Hart YM, Sander JWA, Shorvon SD. National Gen-
eral Practice Study of Epilepsy (NGPSE): partial seizure patterns
sate for the absence of histopathologic confirmation in in a general population. Neurology 1992;42:19 I 1-7.
patients who did not have en bloc resection. However, 4. Spencer SS, Guimaraes P, Katz A, Kim J , Spencer D. Morpho-
these selection criteria limited the analysis to patients logical patterns of seizures recorded intracranially. Epilepsiu 1992;
33 :5 37-45.
with medically intractable epilepsy. It remains unclear to 5. Liiders HO, Engel J Jr, Munari C. General principles. In: Engel J
what extent these observations are applicable to patients Jr. Surgicul treafmenfqfthe epilepsies. 2nd ed. New York: Raven
with less severe HS and MTLE. Another limitation of Press, 1993:137-53.
6. Berkovic SF, Andermann F, Olivier A, et al. Hippocampal sclero-
this study was the small sample size of patients in both sis i n temporal lobe epilepsy demonstrated by magnetic resonance
groups and the small number of interictal discharges in imaging. Ann Neuml 1991;29:175-82.
some patients. Further studies on larger patient popula- 7. Cook M, Fish D, Shorvon S, Straughan K, Stenvens J. Hippocam-
pal volumetric and iiiorphometric studies in frontal and temporal
tions are needed to confirm these results. The patients lobe epilepsy. Brain 1992;115:1001-15.
with tumors in the medial temporal lobe region were 8. Cascino GD. Structural neuroimaging in partial epilepsy: magnetic
highly selected to obtain a well-matched group of pa- resonance imaging. Neurosurg Clin Norrh Am 1995;6:455-64.
9. Wieser H-G, Engel J Jr, Williamson PD, Babb TL, Gloor P. Sur-
tients with disease in the same region as HS patients. gically remediable temporal lobe syndromes. In: Engel J Jr, ed.
Therefore, only patients in whom the tumor was proven Surgical treatment of the epilepsies. 2nd ed. New York: Raven
by MRI and histopathology to be restricted to the medial Press, 1993:49%63.
temporal lobe structures (amygdala a n d o r hippocampus) 10. Engel J Jr, Van Ness PC, Rasmussen TB, Ojemanii LM. Outcome
with respect to epileptic seizures. In: Engel J Jr. Surgical rrearinent
were selected from a much larger group of patients with qf rhe epilepsies. 2nd ed. New York: Raven Press, 1993:609-21.
temporal lobe tumors and epilepsy. However, this re- 11. Gil-Nagel A, Risinger MW. Ictal semiology in hippocampal versus
sulted in a small sample size. extrahippocampal temporal lobe epilepsy. Bruin 1997; 120:183-92.
12. Manford M, Fish DR, Shorvon SD. An analysis of clinical seizure
The establishment of differences between the interictal patterns and their localizing value in frontal and temporal lobe
discharge distributions of the two study groups relied on epilepsies. Brain 1996;l 19:17-40.
closely spaced scalp electrodes according to the 10-10 13. Risinger MW, Engel J Jr, Van Ness PC, Henry TR, Crandall PH.
Ictal localization of temporal lobe seizures with scalphphenoidal
system (1 8) and additional sphenoidal electrodes. The recordings. Neurology 1989;39: 1288-93.
electrodes revealing the most frequent discharge maxi- 14. Gloor P. Preoperative electroencephalographic investigation in
mums were the ipsi- and contralateral sphenoidal elec- temporal lobe epilepsy: extracranial and intracranial recordings.
Can J Neurol Sci 1991;18:S54-8.
trodes followed by the scalp electrodes FT,,,, F,,,, and IS. Williamson PD, French JA, Thadani VM, et al. Characteristics of
F,,,o. This confirms previous studies on the value of medial temporal lobe epilepsy: 11. interictal and ictal scalp elec-
closely spaced electrodes for the accurate localization of troencephalography, neuropsychological testing, neuroimaging,
surgical results, and pathology. Ann Neurol 1993;34:781-7.
epileptiform foci (40,41). The use of a smaller number of
16. Ebner A, Hoppe M. Noninvasive electroencephalography and me-
scalp electrodes (n = 16) without sphenoidal electrodes dial temporal sclerosis. J Clin Neurophysiol 1995;12:23-3 I .
might have contributed to differing findings such as the 17. O’Brien T, Kilpatrick C, Murrie V, Vogrin S, Morris K, Cook MJ.
predominance of posterior temporal discharges in HS Temporal lobe epilepsy caused by medial temporal sclerosis and
temporal neocortical lesions: a clinical and electroencephalo-
patients (17). graphic study of 46 pathologically proven cases. Brain 1996;l 19:
In conclusion, interictal epileptiform discharges in pa- 213341.
tients with MTLE due to HS are highly restricted to the 18. American Electroencephalographic Society. Guidelines in electro-
encephalography, evoked potentials, and polysomnography. J Clin
anterior temporal region with a prevalent involvement of Neurophysiul 1994; 1 1 :1 1 1-3.
sphenoidal electrodes. The underlying HS, rather than 19. Cascino GD, Trenerry MR, So EL, et al. Routine EEG and tem-
the location of the lesion, may be responsible for this poral lobe epilepsy: relation to long-term EEG monitoring, quan-
titative MRI, and operative outcome. Epilep.via I996;37:6S 1-6.
characteristic discharge pattern. In contrast, patients with 20. Klass DW, Westmoreland BS. Nonepileptogenic epileptiform
medial temporal tumors showed more widespread in- electroencephalographic activity. Ann Neuml 1985;I 8:627-35.
terictal discharges. Therefore frequent posterior or extra- 21. Collura TF, Liiders H, Burgess RC. EEG mapping for surgery of
temporal sharp waves may detract from the certainty of epilepsy. Brain Topogr 1990;3:65-77.
22. Watson C, Andermann F, Gloor P, et al. Anatomic basis of amyg-
the diagnosis of HS in complicated cases. Bilateral epi- daloid and hippocampal volume measurement by magnetic reso-
leptiform discharges were not uncommon in both groups. nance imaging. Neurology 1992;42:1743-SO.
23. Cambell MJ, Gardner MJ. Calculation of confidence intervals for
some non-parametric analqses. Br Med J 1988:296: 1454-6.
Acknowledgment: We thank Ed Mascha, M.S., for his as- 24. Quesney LF, Abou-Khalil B, Cole A, Olivier A. Pre-operative
sistance with statistical analyses. extracranial and intracranial EEG investigation in patients with

Eplepsin. Vul. 40, No. 9, IYYY

I268 H. M . HAMER ET AL.

temporal lobe epilepsy: trends, results, and review of pathophysi- 33. Margerison JH, Corsellis JA. Epilepsy and the temporal lobes: a
ologic mechanisms. Acfu Neurol Scand 1988: 117(suppl):52-60. clinical, electroencephalographic and neuropathological study of
25. Ebersole JS, Wade PB. Spike voltage topography identifies two the brain in epilepsy, with particular reference to the temporal
types of frontotemporal epileptic foci. Neurology 1991;4 I : 1425- lobes. Bruin 1966;89:499-530.
33. 34. Sirven JI, Loporace JD, French JA, O’Connor MJ, Sperling MR.
26. Baunigartner C, Lindinger G, Ebner A, et al. Propagation of in- Seizures in temporal lobe epilepsy: 1. Reliability of scalp/
terictal epileptic activity in temporal lobe epilepsy. Neurology sphenoidal ictal recording. Neurology 1997;48:1041-6.
I995;45: 1 18-22. 35. Morrell F. Secondary epileptogenesis in man. Arch Neurol 1985:
27. Spencer DD, Spencer SS, Mattson RH, Williamson PD. Intrace- 42: 3 18-35,
rebral masses in patients with intractable partial epilepsy. Neurol- 36. Holnies MD, Dodrill CV, Ojenxinn GA, Wilensky AJ, Ojemann
ogy 1984;34:432-6. LM. Outcome following surgery in patients with bitemporal in-
28. Awad IA, Rasenfeld J, Ah1 J, Hahn JF, Luders H. tntractable tericral epileptiform patterns. Neurology I997;48: 1037-40.
epilepsy in structural lesions of the brain: mapping, resection strat-
37. Engel J Jr, Rausch R, Lieb 3, Kuhl DE, Crandall PH. Correlation
egies, and seizure outcome. Epilepsia 1991;32: 179-86. of criteria used for localizing epileptic foci in patients considered
29. Boon PA, Williamson PD, Fried I, et al. Intracranial, intraaxial, for surgical therapy of epilepsy. Ann Neurof 1981 ;9:2 15-24.
space-occupying lesions in patients with intractable partial sei-
zures: an anatomoclinical, neuropsychological, and surgical corre- 38. Chnng MY, Walczak TS, Lewis DV, Dawson DV, Radtke R.
lation. Epilepsiu 1991;32:467-76. Temporal lobectomy and independent bitemporal interictal activ-
30. Morris HH, Estes ML, Gilmore R, Van Ness PC, Barnett GH, ity: what degree of lateralization is sufficient? Epilepsia 1991:32:
Turnbull J . Chronic intractable epilepsy as the only symptom of 195-20 1.
primary brain tumor. Epilepsiu 1993;34: 103843. 39. Blume WT, Borghesi JL, Lemieux JF. Interictal indices of tempo-
3 1 . Sammaritano M, de L o t b i d r e A, Andermann F, Olivier A, Gloor ral seizure origin. Ann Neurol 1993:34:703-9.
P, Quesny LF. False lateralization by surface EEG of seizure onset 40. Morris HH, Luders H, Lesser RP, Dinner DS, Klem G.The value
in patients with temporal lobe epilepsy and gross focal cerebral of closely spaced scalp electrodes in the localization of epilepti-
lesions. Ann Neurol 1987;21:361-9. form foci: a study of 26 patients with complex partial seizures.
32. Holmes MD, Dodrill CB, Wilensky AJ, Ojemann LM, Ojemann Electroence~ihulogrClin Neurophysiol 1986;63:107-1 I .
GA. Unilateral focal preponderance of interictal epileptiform dis- 41. Morris HH, Kanner A, Luders H, et al. Can sharp waves localized
charges as a predictor of seizure origin. Arch Neurol 1996:53:228- at the sphenoidal electrode accurately identify a mesio-temporal
32. epileptogenic focus? Epilepsiu 1989;30:532-9.

Epilepsui, Vol. 40, No. 9, I999