braini0113
Brain (1997), 120, 183–192
Ictal semiology in hippocampal versus
extrahippocampal temporal lobe epilepsy
Antonio Gil-Nagel and Michael W. Risinger
Department of Neurology and MINCEP Epilepsy Care, PA,
University of Minnesota, Minneapolis, Minnesota, USA
Summary
We have analysed retrospectively the clinical features and
electroencephalograms in 35 patients with complex partial
seizures of temporal lobe origin who were seizure-free after
epilepsy surgery. Two groups were differentiated for statistical
analysis: 16 patients had hippocampal temporal lobe seizures
(HTS) and 19 patients had extrahippocampal temporal lobe
seizures (ETS) associated with a small tumour of the lateral
or inferior temporal cortex. All patients in the HTS group
had ictal onset verified with intracranial recordings (depth
or subdural electrodes). In the ETS group, extrahippocampal
onset was verified with intracranial recordings in eight
patients and assumed, because of failure of a previous
amygdalohippocampectomy, in one patient. Historical
information, ictal semiology and ictal EEG of typical seizures
were analysed in each patient. The occurrence of early and
late oral automatisms and dystonic posturing of an upper
extremity was analysed separately. A prior history of febrile
Keywords: epilepsy; temporal lobe seizure; epilepsy surgery; mesial temporal sclerosis; aura
Abbreviations: ETS 5 extrahippocampal temporal lobe seizures; HTS 5 hippocampal temporal lobe seizures
Introduction
Temporal lobe seizures may originate in hippocampal and
immediately adjacent structures or arise from extrahippo-
campal neocortical regions. The clinical differentiation of
hippocampal and extrahippocampal temporal lobe epilepsy
has practical value. Hippocampal epilepsy is more prevalent
and is most often associated with hippocampal sclerosis
(Falconer et al., 1964; Dam, 1980; Lieb et al., 1981), a
condition which frequently has an excellent response to
temporal lobectomy if the hippocampus is also removed
(Abou-Khalil et al., 1993). Neocortical temporal lobe epilepsy
is often caused by neoplastic or other space occupying
lesions. Patients with neocortical epilepsy may also benefit
from surgical treatment, but a different surgical approach,
© Oxford University Press 1997
Correspondence to: Antonio Gil-Nagel, MD, Department of
Neurological Sciences, Rush-Presbyterian-St Luke’s
Medical Center, 1653 West Congress Parkway, Chicago,
IL 60612, USA
convulsions was obtained in 13 HTS patients (81.3%) but in
none with ETS (P , 0.0001, Fisher’s exact test). An epigastric
aura preceded seizures in five patients with HTS (31.3%)
and none with ETS (P 5 0.0135, Fisher’s exact test), while
an aura with experiential content was recalled by nine
patients with ETS (47.4%) and none with HTS (P 5 0.0015,
Fisher’s exact test). Early oral automatisms occurred in 11
patients with HTS (68.8%) and in two with ETS (10.5%)
(P 5 0.0005, Fisher’s exact test). Early motor involvement
of the contralateral upper extremity without oral automatisms
occurred in three patients with HTS (18.8%) and in10 with
ETS (52.6%) (P 5 0.0298, Fisher’s exact test). Arrest reaction,
vocalization, speech, facial grimace, postictal cough, late
oral automatisms and late motor involvement of the
contralateral arm and hand occurred with similar frequency
in both groups. These observations show that the early
clinical features of HTS and ETS are different.
partially or completely sparing the hippocampus, is often
undertaken.
The confident localization of temporal lobe seizures
requires a multidisciplinary approach, with utilization of
multiple lines of evidence in most cases. According to
an epilepsy surgery survey (Spencer and Ojeman, 1993),
most epilepsy centres modify the extent of the lateral and
mesial temporal resection based on imaging, invasive
electrophysiological and neuropsychological criteria. When
these tests suggest hippocampal seizure origin, an aggressive
mesial temporal resection is usually undertaken; when
extrahippocampal temporal origin is suspected, a more
conservative hippocampal resection and an extensive
184 A. Gil-Nagel and M. W. Risinger
neocortical resection is preferred. It is possible that certain
aspects of the clinical history, ictal semiology and surface
EEG may help in the differentiation of hippocampal and
extrahippocampal temporal lobe seizures and be used in
conjunction with imaging, electrocorticographic and neuro-
psychological criteria to modify the surgical approach of
medically refractory temporal epilepsy. This study explores
the relationships between hippocampal and neocortical
temporal lobe epilepsy and the clinical history, ictal semiology
and surface ictal EEG of the patients.
Methods
Patients
Cases were selected from 102 consecutive temporal
lobectomies in patients with medically refractory epilepsy
that had intracranial recordings prior to the resection or
removal of a space occupying lesion in a temporal lobe.
The operations were performed from 1986 to 1992 at the
Minnesota Comprehensive Epilepsy Program. All had a
similar evaluation that included interictal EEG, ictal video-
EEG monitoring, neuropsychological testing, intracarotid
amytal test, brain MRI and pathological analysis of the
resected tissue. Prior to surgery, all patients included in the
analysis had one or more seizures per week. Intracranial
recordings were performed in 69 patients (51 without a
tumour and 18 with a tumour). Thirty-three patients with a
tumour in a temporal lobe did not have intracranial recordings
prior to the resection. In order to obtain a pure differentiation
of hippocampal and extrahippocampal temporal lobe seizures,
63 patients were excluded from our analysis. Exclusion was
based on prior history of meningitis, encephalitis or head
injury, conditions associated with a higher risk for multifocal
epilepsy; a tumour .1.5 cm or encroaching upon mesial
temporal structures; and persistence of seizures after the
operation. Of the remaining 39 patients, four were excluded
because ictal intracranial recordings did not demonstrate
distinct hippocampal or extrahippocampal temporal onset.
The 35 patients left are the subjects of the study (Table 1).
After the operation, they continued on antiepileptic
medication and had good outcome, consistent with class I of
Engel’s classification (Engel et al., 1993), for ù1 year. These
patients were divided into two groups. The hippocampal
temporal lobe seizure group (HTS) comprised 16 patients
without space occupying lesions and with a mesial temporal
lobe ictal origin of typical seizures recorded with depth or
subdural electrodes. The extrahippocampal temporal lobe
seizure group (ETS) comprised 19 patients with small tumours
(,1.5 cm in diameter) in the lateral or inferior aspects of
one temporal lobe. At the time of surgery, the average age
was 28.2 years (range 8–38 years) in the HTS group, and
26.7 years (range 16–37 years) in the ETS group.
Clinical variables and follow-up
Follow-up was provided 3, 6 and 12 months after surgery
by the staff at the Minnesota Comprehensive Epilepsy
Program, whenever possible. In some patients who lived
too far away, outcome was assessed by direct telephone
contact. The presence or absence of auras, as well as the
patient’s description, family history, prior history of febrile
convulsions, status epilepticus and generalized tonic–clonic
seizures was obtained from the review of medical records
and direct questioning of the patient in cases of doubt. Auras
were divided according to their content into: (i) epigastric
auras, those associated with a subjective sensation of
discomfort in the epigastric region, with or without nausea
or vomiting; (ii) experiential phenomena consisting of visual
or auditory illusions and hallucinations, complex experiences,
memory flashbacks and illusions of familiarity; (iii) other
auras, including fear, olfactory and gustatory perceptions and
non-lateralizing somatic sensations.
Surgical and histological procedures
Thirteen HTS patients were studied with bilateral depth
electrodes implanted through a posterior and superior
approach. In this procedure, a six-contact point electrode was
stereotactically inserted through the occipitoparietal junction,
longitudinal to the hippocampus. Another four-contact point
electrode array was placed vertically through the parasagittal
frontal lobe, and directed into the posterior orbitofrontal
region. The remaining three HTS patients had two or three
subdural electrodes, of four contact points each, placed
through burr holes in the inferior surface of one temporal
lobe, transversal to the temporal longitudinal axis, with the
most mesial contact point underneath the parahippocampal
gyrus. In the ETS group, eight patients had ictal intracranial
recordings using subdural electrode arrays placed in the
lateral and inferior cortex of one temporal lobe, always
covering the lesion and with two or more contact points
below the parahippocampal gyrus. Another patient in the
ETS group had an unsuccessful selective right amygdalo-
hippocampectomy performed at a different institution 2 years
before being evaluated by us. In the MRI and pathological
analysis, a lateral right temporal astrocytoma was found.
Temporal lobectomies in HTS patients always included
the amygdala, the anterior two-thirds of the hippocampal
formation, the basal and the lateral cortex, sparing the superior
temporal gyrus. The size of the resection varied between
both temporal lobes, with a posterior limit of 5.0–6.0 cm in
the non-dominant hemisphere, and 4.0–4.5 cm in the dominant
side, measured from the anterior wall of the middle cranial
fossa at the time of the operation. Patients in the ETS group
had resections tailored to the lesion, always including a
complete lesionectomy in addition to removal of the anterior
4–6 cm of the temporal lobe. The amygdala and the anterior
one-third of the hippocampus were also removed on occasion,
but mesial resection in the ETS group was never as extensive
as in the HTS group. The lateral and inferior temporal cortex
was resected en bloc, mesial structures were resected in
fragments and via suction aspiration. None of the patients
suffered surgical complications, and postoperative brain MRI
 Semiology of temporal lobe epilepsy 185

Table 1 Clinical data

Patient Febrile Auras Early Late Other ictal Ictal Brain Pathology
no. convulsions OA/MI OA/MI semiology EEG MRI

Patients with temporal lobe epilepsy of hippocampal origin (HTS)

1 Yes – OA MI MA R-HT NL NL
2 Yes Epigastric – OA AR, VO BST NL MTS
3 No Somatic OA1MI – MA, PC BST NL NL
4 Yes Gustatory OA – HA, VO, PC BST NL NL
5 Yes – MI – – R-TT NL NL
6 No Epigastric – MI FG,AR BST NL MTS
7 Yes – MI – FG R-TT NL NL
8 Yes Somatic OA – SP, VO R-TT NL MTS
9 Yes Fear OA – – L-TT NL NL
10 Yes Epigastric OA – MA, VO, PC L-TT L-HCA MTS
11 No Somatic OA1MI – MA, SP R-TT NL NL
12 Yes Fear – – AR, VO, HA L-TT L-HCA MTS
13 Yes Epigastric OA MI VO, SP R-TT NL MTS
14 Yes Epigastric1smell OA1MI – HA, SP BST NL MTS
15 Yes Fear OA1MI – MA R-TT NL NL
16 Yes Smell OA – AR L-HT L-HCA NL
Patients with temporal lobe epilepsy of extrahippocampal origin (ETS)
17 No Voices in a tunnel OA – MA, FG R-TT R-ILT Ganglioglioma
18 No Depersonalization MI – – R-TF R-LT Hamartoma
1dream-like state
19 No Déjà vu – OA AR, VO L-TT L-LT Astrocitoma grade 2
20 No – MI – – R-TT R-LT Ganglioglioma
21 No Somatic – – AR L-TT L-LT Cavernous angioma
22 No – MI OA SP BST R-LT Oligodendroglioma
grade 2
23 No – MI OA – L-TF L-ILT Oligodendroglioma
grade 2
24 No – MI – – R-TF R-IT Oligodendroglioma
grade 1–2
25 No – MI OA VO L-TT L-LT Ganglioglioma
26 No Seeing familiar faces MI OA MA, HA R-TF R-LT Astrocitoma
grade 1–2
27 No Seeing himself from – – AR, MA L-TT L-LT Oligodendroglioma
a distance grade 1
28 No – MI OA VO L-TT L-ILT Astrocitoma
grade 1–2
29 No Déjà vu OA1MI – MA L-TT L-LT Ganglioglioma
30 No Depersonalization – OA HA,MA BST R-LT Astrocitoma
1anxiety grade 1–2
31 No – MI – HA, VO L-TT L-LT Astrocitoma grade 2
32 No – – – FG, MA BST R-LT Ganglioglioma
33 No Somatic MI – FG R-TT R-LT Ganglioglioma
34 No Mirth while – – AR R-TT R-ILT Astrocitoma grade 2
seeing people
35 No Déjà vu – – AR, MA R-TT R-LT Cavernous angioma

MI 5 dystonic motor involvement of an upper extremity; OA 5 oral automatisms; MA 5 manual automatisms; AR 5 arrest reaction; VO 5
vocalization; SP 5 speech; HA 5 hyperactivity; FG 5 facial grimace; PC 5 postictal cough; R- 5 right; L- 5 left; BST 5 bilateral
synchronous theta; HT 5 hemispheric theta; TT 5 temporal theta; TF 5 temporal fast; HCA 5 hippocampal atrophy; NL 5 normal; ILT 5
inferior lateral temporal tumor; LT 5 lateral temporal tumor; IT 5 inferior temporal tumor; MTS 5 mesial temporal sclerosis.

did not demonstrate areas of infarction in the posterior margin
of the resection. Specimens were fixed in 10% formalin,
dehydrated with ethanol and processed with toluidine and
paraffin. Sections of 5 µm were then obtained along the
longitudinal axis of the temporal lobe, and from mesial
temporal fragments. Sections were stained with haematoxylin
and eosin, and additional special stains to demonstrate tumour
differentiation were performed as required.
Ictal behaviour
A total of 289 seizures were reviewed (this included all video
ictal recordings in each patient) to verify that only one type
of complex partial seizure was present. Patients with multiple
seizure types were excluded from the study, except those in
whom different seizures were the result of different degrees
of propagation of a hypothetical single epileptic focus (i.e.
simple partial, complex partial and secondarily generalized
186 A. Gil-Nagel and M. W. Risinger
tonic–clonic seizures). The number of seizures recorded
averaged 7.6 (range 5–13) in the HTS group and 8.8 (range
3–16) in the ETS group. A typical video ictal recording, with
good quality surface EEG and appropriate view of the face,
trunk and upper extremities, was selected for sequential
analysis. We noted the presence or absence of the following
clinical ictal behaviours: (i) oral automatisms including
chewing, lip smacking, tongue protruding and lip pursing;
(ii) dystonic motor involvement of an upper extremity; (iii)
facial grimace or tonic contraction of the muscles of facial
expression; (iv) total arrest reaction ù20 s; (v) vocalization,
including screaming and grunting; (vi) preserved speech,
defined as saying two or more words during the seizure;
(vii) manual automatisms, including manual exploratory
behaviour, grabbing and rubbing; (viii) hyperactivity; (ix)
postictal cough. Because of their independence from external
stimuli, oral automatisms and motor involvement provide a
temporal resolution that allowed analysis in two intervals:
(i) patients were considered to have early oral automatisms
and early motor involvement, when these clinical features
started in the first 20 s, counted from the first ictal change
on surface EEG; (ii) late oral automatisms and late motor
involvement were considered when these features started
.20 s after the first ictal EEG change. This distinction was
not possible with other ictal behaviour because its occurrence
was frequently triggered, or limited, by environmental
stimulation (e.g. manual automatisms will start when an
object is placed in the patient’s hand or visual field, speech
will occur when the patient is questioned and arrest will
cease when someone interacts with the patient).
Electroencephalogram
Surface EEGs had been performed with gold cup electrodes
placed according to the International 10/20 System applied
by collodion, and sphenoidal electrodes inserted through the
mandibular notch, linked in bipolar and referential montages.
All ictal EEG recordings of each patient were reviewed
initially to ascertain the presence of one single seizure type.
From this initial pool, three good quality recordings were
selected in each patient to study the ictal patterns. The EEG
changes considered as ictal onset included diffuse or focal
slowing, spikes or fast activity. But, only four patterns were
identified: (i) bilateral synchronous theta, where diffuse
rhythmic theta slowing was recorded during the ictal onset;
(ii) ipsilateral hemispheric theta, in which non-localizing
unilateral rhythmic theta slowing was observed on the side
of the resected temporal lobe; (iii) ipsilateral temporal theta,
where sphenoidal or anterior temporal (T3 or T4) maximum
sharp theta recruiting rhythm was recorded in the side of the
operated temporal lobe; and (iv) ipsilateral temporal fast, in
which focal 8 Hz or faster frequency evolving to a recruiting
rhythm was present in the operated temporal region.
Statistical analysis
Either χ2 analysis or Fisher’s exact test was used to evaluate
the significance of a relationship of specific clinical features,
Table 2 Results of intracranial recordings in 105 seizures
recorded in 28 patients who had complete seizure control
after temporal lobectomy.
n: HTS (%) n: ETS (%)
Number of patients 19 9
Total number of seizures 71 34
Seizures with hippocampal onset 62 (87.3) 0
Seizures with neocortical onset 0 32 (94.1)
Seizures with simultaneous 6 (8.5) 2 (5.9)
hippocampal and neocortical onset
Seizures with non-localizing onset 3 (4.2) 0
Patients with prior history of meningitis, encephalitis or head
trauma are excluded. The hippocampal temporal seizure (HTS)
group includes patients without a space-occupying lesion in brain
MRI. Extrahippocampal temporal seizure (ETS) group comprises
patients with a tumour ,1.5 cm in diameter in the temporal
neocortex. Numbers and percentages of seizures are shown.
ictal semiology and surface EEG findings to the focus of
ictal origin, assuming statistical significance when P , 0.05.
Results
Intracranial recordings
A total of 105 seizures recorded with intracranial electrodes
were analysed (Table 2). This included all intracranial ictal
recordings in 28 patients, 19 in the HTS group and nine in
the ETS group. The number of seizures recorded averaged
3.1 per patient (range 2–6) in the HTS group and 2.66 (range
2–5) in the ETS group. Seizure onset was defined, according
to previously described ictal patterns (Spencer et al., 1990,
1992), as a localized recruiting rhythm faster than 2 Hz or
rhythmic spiking with a frequency of 0.5–2 Hz. Three patients
in the HTS group and one in the ETS group were excluded
from the analysis of ictal semiology because adequate
localization of ictal onset was not possible. One HTS patient,
who had two seizures recorded, was excluded because both
seizures had simultaneous hippocampal and lateral temporal
onset. Another HTS patient, who had three seizures recorded,
was excluded because two seizures had simultaneous
hippocampal and lateral onset and one seizure was non-
localizing. The third HTS patient excluded had two seizures
recorded, one with simultaneous mesial and lateral onset
and another with non-localizing onset. Two more seizures in
the HTS group had simultaneous hippocampal and
extrahippocampal onset, each of them was recorded in a
different patient. This included one patient who had four
more seizures with mesial onset and another patient who had
three additional mesial onset seizures. Because the majority
of the seizures in these two HTS patients were of mesial
temporal origin, they were not excluded from the study, and
ictal semiology was analysed in seizures with proven mesial
onset. None of the seizures in the HTS group had
extratemporal onset. One ETS patient with a small low grade
glioma in the inferior lateral temporal lobe was excluded
because his two recorded seizures had ictal onset

Table 3 Differences in historical information and auras between HTS and ETS patients
Clinical feature
Number of patients
History of febrile convulsions
History of generalized tonic–clonic seizures
History of status epilepticus
Family history of epilepsy†
Presence of aura
Epigastric aura
Olfactory and gustatory aura
Aura of fear
Non-lateralizing somatic aura‡
Experiential phenomena
Numbers and percentages of HTS and ETS patients with each feature are shown. †Includes family history of epilepsy in first or second
degree relatives. ‡Includes auras of non-lateralizing somatic sensation, excluding epigastric sensations. §Apart from the two χ2 tests
indicated, Fisher’s exact test was used. *Statistically significant.
simultaneously in the mesial and lateral temporal electrodes.
None of the seizures in ETS patients had exclusively mesial,
extratemporal or non-localizing onset.
Clinical history
Surgery was performed in the right temporal lobe in 21
patients (10 HTS, 11 ETS). There was a history of febrile
convulsions (Table 3) in 13 patients with HTS (81.3%)
compared with none in the ETS group (P , 0.0001, Fisher’s
exact test). In 10 of these patients, the onset of afebrile
seizures occurred after a seizure-free period averaging 8.8
years (range 4–17 years). Mean age at the onset of afebrile
seizures was similar in both groups: 12 years (range 5–19
years) for HTS patients and 15.2 years (range 3–25 years)
for ETS patients. Age at the time of surgery was also similar
in both groups: 28.2 years (range 8–41 years) in HTS patients
and 26.7 years (range 16–37 years) in ETS patients. Prior
history of status epilepticus was obtained in four HTS patients
(25%) and two ETS patients (10.5%), a difference which
was not statistically significant. In the HTS group, 10 patients
(62.5%) reported having had one or more generalized tonic–
clonic seizures in their lifetime, compared with 12 in the
ETS group (63.1%), also a non-significant difference.
Auras
Epigastric auras (Table 3) were present in five patients with
HTS (31.3%), and none with ETS (P 5 0.0135, Fisher’s
exact test). One of these patients also perceived a peculiar
foul smell at the onset of her seizures. Experiential phenomena
occurred exclusively in patients from the ETS group, in
which nine patients (47.4%) described them as their typical
aura (P 5 0.0015, Fisher’s exact test). Of these patients
with experiential auras, three had ‘déjà vu’ sensations, two
described them as a dream associated with a feeling of
depersonalization, two reported seeing people or faces which
evoked a feeling of familiarity in one of them, and mirth in
the other, one patient described autoscopic phenomena and
Semiology of temporal lobe epilepsy 187
n: HTS (%) n: ETS (%) P-value§
16 19
13 (81.3) 0 0.0001*
10 (62.5) 12 (63.2) 0.9680 (χ2)
4 (25.0) 2 (10.5) 0.3791
2 (12.5) 5 (26.3) 0.4150
13 (81.3) 11 (57.9) 0.1382 (χ2)
5 (31.3) 0 0.0135*
3 (18.8) 0 0.0856
3 (18.8) 0 0.0856
3 (18.8) 2 (10.5) 0.6418
0 9 (47.4) 0.0015*
another reported finding himself in a tunnel where he could
hear strange voices (Table 1). Auras producing an illusion
of foul smell, abnormal taste or fear were present exclusively
in the HTS group, but their occurrence was rare in this study,
and does not differ significantly among ETS and HTS
patients. Non-localizing or midline somatic sensations were
almost equally present in each patient group.
Ictal behaviour
We found significant differences in early clinical behaviour
when comparing seizures of hippocampal and extrahippo-
campal origin (Table 4). Early oral automatisms were present
in 11 patients with HTS (68.8%), but only in two patients
with ETS (10.5%) (P 5 0.0005, Fisher’s exact test). In two
cases, both in the HTS group, oral automatisms were the
only clinical manifestation during the first 20 s from seizure
onset. Five had oral automatisms associated with motor
involvement of an upper extremity. There was no difference
in the occurrence of early motor involvement between both
groups when including patients who also had early oral
automatisms. However, early motor involvement without
early oral automatisms was recorded in only two patients
with HTS (12.5%) compared with 10 with ETS (52.6%)
(P 5 0.0298, Fisher’s exact test). Late oral automatisms
occurred in one patient with HTS and seven with ETS;
therefore, when counted together, early and late oral
automatisms were found in 12 patients with HTS (75%) and
nine of the ETS patients (47.4%), an insignificant difference
(χ2 5 2.763; P , 0.10). Late motor involvement occurred
in three HTS patients (18.8%) but not in ETS patients;
together, early and late motor involvement was recorded in
nine HTS patients (56.3%) and 11 ETS patients (57.9%),
also not significantly different (χ2 5 0.10; P , 0.92). No
difference was found between both groups in the occurrence
of arrest reaction, facial grimace, hyperactivity, manual auto-
matisms, vocalization and speech. Ictal speech occurred in
four patients of the HTS group (25%) and one of the ETS
group (P 5 0.1558, Fisher’s exact test). All these were
188 A. Gil-Nagel and M. W. Risinger

Table 4 Ictal behaviour in hippocampal and extrahippocampal temporal lobe seizures

Ictal behaviour n: HTS (%) n: ETS (%) P-value†

Number of patients 16 19
Early oral automatisms 11 (68.8) 2 (10.5) 0.0005*
Early and late oral automatisms 12 (75.0) 9 (47.4) 0.0965 (χ2)
Early motor IUE 6 (37.6) 11 (57.9) 0.3001 (χ2)
Early motor IUE without oral automatisms 2 (12.5) 10 (52.6) 0.0298*
Early and late motor IUE 9 (56.3) 11 (57.9) 0.9220 (χ2)
Manual automatisms 5 (31.3) 7 (36.8) 0.7284 (χ2)
Arrest reaction 4 (25.0) 5 (26.3) .0.9999
Hyperactivity 3 (18.8) 3 (15.8) .0.9999
Speech 4 (25.0) 1 (5.3) 0.1558
Vocalization 6 (37.5) 4 (21.1) 0.4543
Face grimace 2 (12.5) 3 (15.8) .0.9999
Postictal cough 3 (18.8) 0 0.0856

Numbers and percentages of HTS and ETS patients with each behaviour are shown. IUE 5 invovement of upper extremity. †Apart from
the four χ2 tests indicated, Fisher’s exact test was used. *Statistically significant.

Table 5 Ictal EEG in hippocampal and extrahippocampal temporal lobe seizures

EEG pattern n: HTS (%) n: ETS (%) P-value†

Number of patients 16 19
Bilateral synchronous theta 5 (31.3) 3 (15.8) 0.4236
Ipsilateral hemispheric rhythmic theta 2 (12.5) 0 0.2017
Ipsilateral temporal rhythmic theta 9 (56.3) 12 (63.2) 0.7391
Ipsilateral temporal fast 0 4 (21.1) 0.1088

Numbers and percentages of HTS and ETS patients with each pattern are shown. †Fisher’s exact test was used.

patients with seizure onset in the non-dominant temporal
lobe. No difference was demonstrated when accounting only
for right temporal lobe ictal onset: four of ten right temporal
lobe patients in the HTS group (40%) versus one of 11 right
temporal lobe patients in the ETS group (P 5 0.1486,
Fisher’s exact test). Although postictal cough occurred only
in patients of the HTS group, the number (three patients or
18.8%) was too low to reach statistical significance.
Ictal EEG
We found no difference between the two groups in the
occurrence of the four different EEG patterns described
(Table 5). The most common pattern was ipsilateral temporal
rhythmic theta which was encountered in 21 patients (60%);
nine from the HTS group (56.3%) and 12 in the ETS group
(63.2%). The second most common pattern was bilateral
synchronous theta, present in eight patients (22.9%); five
from the HTS group (31.3%) and three from the ETS group
(15.8%). Ipsilateral temporal fast activity was present only
in the ETS group (four patients or 21.1%), but this was not
statistically significant (P 5 0.1088, Fisher’s exact test).
Ipsilateral hemispheric rhythmic theta was only present in
two patients, both from the HTS group.
Intrinsic differences within the ETS and HTS
groups
No difference in the occurrence of experiential aura and early
motor involvement was detected when comparing patients in
the ETS group with proven extrahippocampal onset (eight
studied with intracranial recordings and one who failed a
previous selective amygdalohippocampectomy) versus 10
with assumed ETS onset on the basis of their small neocortical
lesion and excellent response to surgery. In the HTS group,
we found no difference in the occurrence of epigastric auras
and early oral automatisms when comparing the eight patients
with histologically proven mesial temporal sclerosis with
eight without pathological diagnosis.
Discussion
The confident differentiation of hippocampal and extra-
hippocampal seizures (with non-invasive evaluation) requires
consideration of several lines of evidence. The presence of
hippocampal atrophy is a strong but not perfect predictor of
mesial temporal ictal origin. Likewise, the presence of a
discrete extrahippocampal structural lesion is strong pre-
sumptive evidence of extrahippocampal ictal origin. Extra-
cranial electrophysiological measurements may suggest a

focus of ictal origin, but the specificity of these measures is
unknown (Morris et al., 1989; Risinger et al., 1989; Ebersole
and Wade, 1991). 99 m technetium hexamethyl propylene-
amine-oxime–single photon emission computed tomography
may demonstrate hypoperfusion interictally and hyperper-
fusion ictally, but it usually involves both the mesial and
lateral aspects of the affected temporal lobe, and in the
immediate postictal analysis it is common to observe mesial
hyperperfusion and lateral hypoperfusion regardless of mesial
or lateral onset (Rowe et al., 1991). Hypometabolism patterns
detected with interictal [18F]fluorodeoxyglucose PET usually
do not distinguish between hippocampal and extrahippo-
campal temporal lobe epilepsy (Henry et al., 1991), except
in those cases associated with a lateral structural abnormality
(Hajek et al., 1993). Many experts now believe that some
patients with hippocampal seizures may be identified by
recognizing the presence of the electroclinical syndrome of
mesial temporal lobe epilepsy (Engel, 1992; French et al.,
1993). This syndrome consists of characteristic electrophysio-
logical, structural and clinical features; our preselected popu-
lation of patients with hippocampal ictal origin comprised
eight patients for whom this syndromic diagnosis was demon-
strated in the pathological analysis and eight more patients
without such confirmation. Since we found no difference in
the occurrence of febrile convulsions, epigastric aura and
early oral automatisms in either group, we concluded that a
larger number of HTS patients may have mesial temporal
sclerosis, but this could not be detected because of a sample
bias when the hippocampal formation is removed via suction.
Our findings document a statistically significant association
of history of febrile convulsions in early childhood, epigastric
aura and early oral automatistic ictal behaviour with a mesial
temporal seizure origin. Our study also indicates that the
presence of an aura with experiential phenomena and early
unilateral somatic motor involvement without accompanying
oral automatistic behaviour is significantly associated with a
neocortical temporal ictal origin. Visceral sensations are the
most common aura in patients with hippocampal seizure
origin (Wieser and Williamson, 1993; Fried et al., 1995;
Kotagal et al., 1995), and they can be elicited by electrical
stimulation of the mesial temporal lobe (Penfield and Perot,
1963). Our study does not differ from these findings, and
provides further support of their originating from the
amygdala and hippocampal formation when occurring at the
onset of a seizure. In our study, we found no lateralization
value of experiential phenomena and epigastric auras. This
is in agreement with the results of Palmini and Gloor (1992)
and Fried et al. (1995) and disagrees with the findings of
Gupta et al. (1983) and Taylor and Lochery (1987). There is
some controversy about a possible mesial versus lateral
temporal origin of experiential auras (Blume et al., 1993;
Wieser and Williamson, 1993). Penfield was unable to elicit
these phenomena from stimulation of the hippocampal
formation (Penfield and Perot, 1963). Gloor studied 10
patients with depth electrodes who had fear and experiential
phenomena at onset of their seizures. He was able to trigger
Semiology of temporal lobe epilepsy 189
the aura by electrical stimulation in six of them. In most
cases both the mesial and neocortical structures were involved
in the discharge. The second most common pattern was
involvement of the limbic structures without propagation to
lateral electrodes, and only on a few occasions did the
discharge selectively involve lateral contact points of depth
electrodes (Gloor et al., 1982). This study, because of the
lack of structural abnormalities and the use of depth electrodes
orthogonally placed in the hippocampal formations, was
limited by a relatively small and random sampling of the
temporal neocortex compared with the hippocampal sample.
In our study, intracranial recordings in the ETS group were
guided by the presence of a structural abnormality that limited
the area of possible ictal onset within the relatively large
neocortical temporal cortex. In addition, in this study, we
separate auras with fear from other experiential auras, and
found fear auras to occur only in patients with hippocampal
seizures, indicating that experiential auras and fear auras
may have a different anatomical substrate, with the former
occurring from activation of temporal neocortex and the
latter from the amygdala, hippocampus and parahippocampal
formation. Our findings may support Gloor’s hypothesis that
experiential phenomena result from the activation of neurons
interconnected in a matrix that includes elements from the
hippocampal formation as well as the temporal isocortex
(Gloor, 1990). In this model, a discharge originating in
the neocortex will interact with a functioning hippocampal
correlate; therefore, chances of becoming conscious of
experiential phenomena will be higher. However, a discharge
originating in the hippocampal formation will most likely
continue propagation in the hippocampus and provoke its
paralysis before its neocortical correlate is activated. In
this case, the synchronous functioning of the hippo-
campal–neocortical neural matrix is less likely.
The temporal resolution of ictal behaviour, as revealed by
video recordings, has been used in numerous studies to
identify subgroups of temporal lobe epilepsies (Wieser, 1983;
Walsh and Delgado-Escueta, 1984; Delgado-Escueta and
Walsh, 1985; Maldonado et al., 1988; Kotagal et al., 1995).
This has been valuable in identifying the basic mechanisms
of ictal propagation and the anatomical substrate of some
ictal features. However, because of their complexity, some
of these approaches are difficult to use in clinical practice.
Other authors found differences in ictal semiology when
comparing patients with temporal lobe tumours in any
location versus patients with mesial temporal sclerosis (Saygi
et al., 1994), but because outcome after surgery and
intracranial recordings were not part of their patient selection,
the use of ictal semiology for distinction between HTS and
ETS is limited. In this study, we designed a simple method
to time the occurrence of some ictal features (i.e. the
occurrence of motor involvement or oral automatisms within
20 s from the first ictal change in the EEG), and found that
it is useful in discriminating seizures of hippocampal onset
from extrahippocampal temporal seizures. We also found that
the presence of the same features later in the seizure does not
190 A. Gil-Nagel and M. W. Risinger
differ between the HTS and ETS groups. This semiological
difference suggests different propagation patterns in
hippocampal and extrahippocampal temporal lobe epilepsy.
Oroalimentary automatisms occur either as a release
phenomenon or, less likely, as a positive manifestation due
to the abnormal discharge propagating to the limbic system
(Penfield, 1955). Neocortical temporal lobe seizures will
preferentially invade the basal ganglia (Newton et al., 1992)
or the frontal cortex (Wieser et al., 1993) or both, inducing
predominantly dystonic activity of the contralateral upper
extremity. However, differences in the occurrence of oral
automatisms and motor involvement among these two
subtypes of temporal lobe epilepsy are not encountered when
the same variables are analysed without considering their
time of appearance during the seizure. This may be related
to the existence of common final pathways of propagation
when the ictal discharge reaches a critical extension, and
stresses the need to analyse early semiological features.
Our patient populations were preselected to approximate
a pure differentiation of hippocampal and extrahippocampal
cases. This was achieved by excluding patients with risk
factors for multifocal epilepsy, lack of complete seizure
control after surgery or poor localization with intracranial
recordings. However, some ambiguity still exists in our
selection process. In particular, it is possible that some of
our patients with extrahippocampal localization may represent
cases of dual pathology, with hippocampal sclerosis in
addition to the documented extrahippocampal lesion. This is
not a major concern in our study, however, because brain
MRI did not demonstrate hippocampal atrophy or abnormal
signal in the ETS group and the lesions in question are
almost all glial tumours (Table 1), lesions that have a low
association with additional hippocampal pathology (Lévesque
et al., 1991; Nakasato et al., 1992; Saygi et al., 1994). Only
one patient in this series had a hamartoma, and he also
experienced early dystonic posturing of an upper extremity,
thus, not separating from the most common ictal behaviour
in the ETS group. We had difficulty assuming that a structural
abnormality and complete seizure control after its removal
is proof of ictal localization, although that is accepted by
most authors (Wieser and Williamson, 1993; Burgerman
et al., 1995; Walczak, 1995). For this reason, we compared
two categories of patients within the ETS group; those with
ictal localization supported by the presence of a lesion and
seizure resolution following surgery, and those who, in
addition, had ictal onset demonstrated with intracranial
recordings, or, in one case, failure of surgery when it did not
include the temporal neocortex. The two ETS subgroups did
not differ in the occurrence of experiential auras and early
ictal behaviour. Furthermore, in ETS patients studied with
intracranial electrodes, after excluding patients with
hippocampal and parahippocampal tumours, large tumours
and seizure persistence after surgery, 94% of seizures had
exclusive neocortical onset (Table 2), indicating that
significant variation between the two ETS subgroups is
very unlikely.
In this study, results of intracranial recordings were used
as part of the selection process. In the HTS group the majority
of seizures had exclusive hippocampal onset, probably
because patients with multifocal pathology and less than
excellent surgical outcome had already been excluded from
the analysis. Some patients in the ETS group were studied
with inferior and lateral temporal subdural electrodes, a
technique that establishes precisely the epileptogenic focus
in most patients with neocortical temporal epilepsy (Lüders
et al., 1989; Wyler, 1991). Apart from one case, that was
excluded from the analysis of semiology, ETS patients had
ictal onset in the inferior lateral or lateral temporal electrodes
before involvement of mesial electrodes, a pattern of
propagation that excludes hippocampal origin. These results
are not different from other observations (Spencer et al.,
1990). Finally, in our study, ictal EEG was not helpful in the
differentiation of HTS versus ETS. This may be related, in
part, to our restrictive selection criteria, which excluded
patients that did not need intracranial recordings because
other lines of evidence, including ictal EEG, were sufficient
to localize their seizure onset. Had these patients been
included in the study, the incidence of focal fast frequency
and rhythmic focal theta would have been higher and may
have resulted in some discrimination value between HTS
and ETS.
Our results indicate that early clinical ictal characteristics
and EEG cannot be used in isolation to distinguish
hippocampal from extrahippocampal localization. But some
features of early ictal semiology, such as the content of auras
and the presence of oral automatisms and dystonic posturing
of an upper extremity, do have significant associations with
limbic versus neocortical temporal localization. When
evaluating a patient with presumed temporal lobe epilepsy
for surgical treatment, a meticulous analysis of early ictal
behaviour can provide valuable adjunctive information.
Acknowledgements
The research was supported in part by grants FIS 91/5576 and
93/5077, Spain (A.G.-N.) and NIH.grant P50-NS16308, USA.
References
Abou-Khalil B, Andermann E, Andermann F, Olivier A, Quesney
LF. Temporal lobe epilepsy after prolonged febrile convulsions:
excellent outcome after surgical treatment. Epilepsia 1993; 34:
878–83.
Blume WT, Girvin JP, Stenerson P. Temporal neocortical role in
ictal experiential phenomena. Ann Neurol 1993; 33: 105–7.
Burgerman RS, Sperling MR, French JA, Saykin AJ, O’Connor
MJ. Comparison of mesial versus neocortical onset temporal lobe
seizures: neurodiagnostic findings and surgical outcome. Epilepsia
1995; 36: 662–70.
Dam AM. Epilepsy and neuron loss in the hippocampus. Epilepsia
1980; 21: 617–29.

Delgado-Escueta AV, Walsh GO. Type I complex partial seizures of
hippocampal origin: excellent results of anterior temporal lobectomy.
Neurology 1985; 35: 143–54.
Ebersole JS, Wade PD. Spike voltage topography identifies two
types of frontotemporal epileptic foci. Neurology 1991; 41: 1425–33.
Engel J Jr. Update on surgical treatment of the epilepsies. Summary
of the Second International Palm Desert Conference on the Surgical
Treatment of the Epilepsies (1992). Neurology 1993; 43: 1612–7.
Engel J Jr, Van Ness PC, Rasmussen TB, Ojemann LM. Outcome
with respect to epileptic seizures. In Engel J Jr, editor. Surgical
treatment of the epilepsies. 2nd ed. New York: Raven Press, 1993:
609–21.
Falconer MA, Serafetinides EA, Corselis JA. Etiology and
pathogenesis of temporal lobe epilepsy. Arch Neurol 1964; 10:
233–48.
French JA, Williamson PD, Thadani VM, Darcey TM, Mattson RH,
Spencer SS, et al. Characteristics of medial temporal lobe epilepsy:
I. Results of history and physical examination. Ann Neurol 1993;
34: 774–80.
Fried I, Spencer DD, Spencer SS. The anatomy of epileptic auras:
focal pathology and surgical outcome. J Neurosurg 1995; 83: 60–6.
Gloor P. Experiential phenomena of temporal lobe epilepsy. Facts
and hypotheses.[Review]. Brain 1990; 113: 1673–94.
Gloor P, Olivier A, Quesney LF, Andermann F, Horowitz S. The
role of the limbic system in experiential phenomena of temporal
lobe epilepsy. Ann Neurol 1982; 12: 129–44.
Gupta AK, Jeavons PM, Hughes RC, Covanis A. Aura in temporal
lobe epilepsy: clinical and electroencephalographic correlation. J
Neurol Neurosurg Psychiatry 1983; 46: 1079–83.
Hajek M, Antonini A, Leenders KL, Wieser HG. Mesiobasal versus
lateral temporal lobe epilepsy: metabolic differences in the temporal
lobe shown by interictal 18F-FDG positron emission tomography
[see comments]. Neurology 1993; 43: 79–86. Comment in:
Neurology 1993; 43: 2156
Henry TR, Sutherling WW, Engel J Jr, Risinger MW, Lévesque
MF, Mazziotta JC, et al. Interictal cerebral metabolism in partial
epilepsies of neocortical origin. Epilepsy Res 1991; 10: 174–82.
Kotagal P, Lüders HO, Williams G, Nichols TR, McPherson J.
Psychomotor seizures of temporal lobe onset: analysis of symptom
clusters and sequences. Epilepsy Res 1995; 20: 49–67.
Lévesque ML, Nakasato N, Vinters HV, Babb TL. Surgical treatment
of limbic epilepsy associated with extrahippocampal lesions: the
problem of dual pathology. J Neurosurg 1991; 75: 364–70.
Lieb JP, Engel J Jr, Brown WJ, Gevins AS, Crandall PH.
Neuropathological findings following temporal lobectomy related
to surface and deep EEG patterns. Epilepsia 1981; 22: 539–49.
Lüders H, Hahn J, Lesser RP, Dinner DS, Morris HH 3d, Wyllie
E, et al. Basal temporal subdural electrodes in the evaluation of
patients with intractable epilepsy. Epilepsia 1989; 30: 131–42.
Maldonado HM, Delgado-Escueta AV, Walsh GO, Swartz BE, Rand
RW. Complex partial seizures of hippocampal and amygdalar origin.
Epilepsia 1988; 29: 420–33.
Semiology of temporal lobe epilepsy 191
Morris HH 3d, Kanner A, Lüders H, Murphy D, Dinner DS, Wyllie
E, et al. Can sharp waves localized at the sphenoidal electrode
accurately identify a mesio-temporal epileptogenic focus? Epilepsia
1989; 30: 532–9.
Nakasato N, Lévesque MF, Babb TL. Seizure outcome following
standard temporal lobectomy: correlation with hippocampal neuron
loss and extrahippocampal pathology. J Neurosurg 1992; 77: 194–
200.
Newton MR, Berkovic SF, Austin MC, Reutens DC, McKay WJ,
Bladin PF. Dystonia, clinical lateralization, and regional blood flow
changes in temporal lobe seizures. Neurology 1992; 42: 371–7.
Palmini A, Gloor P. The localizing value of auras in partial seizures:
a prospective and retrospective study. Neurology 1992; 42: 801–8.
Penfield W. The twenty-ninth Maudsley Lecture: the role of the
temporal cortex in certain psychical phenomena. J Ment Sci 1955;
101: 451–65.
Penfield W, Perot P. The brain’s record of auditory and visual
experience: a final summary and discussion. Brain 1963; 86:
595–620.
Risinger MW, Engel J Jr, Van Ness PC, Henry TR, Crandall PH.
Ictal localization in temporal lobe seizures with scalp/sphenoidal
recordings. Neurology 1989; 39: 1288–93.
Rowe CC, Berkovic SF, Austin MC, McKay WJ, Bladin PF. Patterns
of postictal cerebral blood flow in temporal lobe epilepsy: qualitative
and quantitative analysis. Neurology 1991; 41: 1096–103.
Saygi S, Spencer SS, Scheyer R, Katz A, Mattson R, Spencer DD.
Differentiation of temporal lobe ictal behavior associated with
hippocampal sclerosis and tumors of temporal lobe. Epilepsia 1994;
35: 737–42.
Spencer DD, Ojemann GA. Overview of therapeutic procedures.
In: Engel J Jr, editor. Surgical treatment of the epilepsies. 2nd ed.
New York: Raven Press, 1993: 455–71.
Spencer SS, Spencer DD, Williamson PD, Mattson R. Combined
depth and subdural electrode investigation in uncontrolled epilepsy.
Neurology 1990; 40: 74–9.
Spencer SS, Guimaraes P, Katz A, Kim J, Spencer DD.
Morphological patterns of seizures recorded intracranially. Epilepsia
1992; 33: 537–45.
Taylor DC, Lochery M. Temporal lobe epilepsy: origin and
significance of simple and complex auras. J Neurol Neurosurg
Psychiatry 1987; 50: 673–81.
Walczak TS. Neocortical temporal lobe epilepsy: characterizing the
syndrome [editorial]. Epilepsia 1995; 36: 633–5.
Walsh GO, Delgado-Escueta AV. Type II complex partial seizures:
poor results of anterior temporal lobectomy. Neurology 1984; 34:
1–13.
Wieser HG. Discussion of 5 typical seizure constellations in the
light of neuroanatomy. In: Wiesser HG. Electroclinical features
of the psychomotor seizure. Stuttgart-London: Gustav Fischer-
Butterworths, 1983: 193–208.
Wieser HG, Williamson PD. Ictal semiology. In: Engel J Jr, editor.
Surgical treatment of the epilepsies. 2nd ed. New York: Raven
Press, 1993: 161–71.
192 A. Gil-Nagel and M. W. Risinger

Wieser HG, Engel J Jr, Williamson PD, Babb TL, Gloor P. Surgically
remediable temporal lobe syndromes. In: Engel J Jr, editor. Surgical
treatment of the epilepsies. 2nd ed. New York: Raven Press, 1993:
49–63.
Wyler AR. Subdural strip electrodes in surgery of epilepsy. In:
Lüders H, editor. Epilepsy surgery. New York: Raven Press, 1991:
395–8.
Received August 6, 1996. Accepted October 11, 1996