Clinical significance of ictal high frequency oscillations in medial temporal lobe epilepsy.

Abstract

To clarify the clinical significance of ictal high frequency oscillations (HFO) in the medial temporal lobe.
This study included 19 patients who underwent intracranial electrode implantation in bilateral temporal lobes and had at least one seizure recorded at 1kHz sampling rate. The characteristics of ictal HFO in the medial temporal lobe, and the relations between the presence of HFO, pathology, and postoperative seizure outcome were analyzed.
Ictal HFO were detected from medial temporal structures in 11 patients with medial temporal lobe epilepsy (MTLE). Among eight patients without HFO, only three were diagnosed with MTLE. Ictal HFO were detected from unilateral medial temporal structures ipsilateral to the side of hippocampal sclerosis (HS). In one patient with bitemporal independent seizure onset, ictal HFO were detected only on the side of HS. HS was detected in all 11 patients with HFO, but in only one of four patients without HFO. Seizure outcome did not differ between patients with and without HFO.
Ictal HFO in the medial temporal lobe may be a specific marker for MTLE with HS.
Recording of ictal HFO in the medial temporal lobe may be useful for presurgical evaluation of MTLE.

Full text

Page 1

Clinical significance of ictal high frequency oscillations in medial temporal
lobe epilepsy
Naotaka Usui⇑, Kiyohito Terada, Koichi Baba, Kazumi Matsuda, Fumihiro Nakamura, Keiko Usui,
Miyako Yamaguchi, Takayasu Tottori, Shuichi Umeoka, Shigeru Fujitani, Akihiko Kondo,
Tadahiro Mihara, Yushi Inoue
National Epilepsy Center, Shizuoka Institute of Epilepsy and Neurological Disorders, Shizuoka, Japan
a r t i c l e i n f o
Article history:
Accepted 7 February 2011
Available online 12 March 2011
Keywords:
High frequency oscillations
Medial temporal lobe epilepsy
Intracranial EEG
Fast ripples
Epilepsy surgery
h i g h l i g h t s
? Ictal high frequency oscillations (HFO) were detected unilaterally.
? They were detected ipsilateral to the side of hippocampal sclerosis (HS).
? They were not propagated contralaterally except for one patient.
? In one case with bitemporal onset, ictal HFO were detected only on the side of HS.
? Ictal HFO in the medial temporal lobe is the electrophysiological signature of HS.
a b s t r a c t
Objective: To clarify the clinical significance of ictal high frequency oscillations (HFO) in the medial tem-
poral lobe.
Methods: This study included 19 patients who underwent intracranial electrode implantation in bilateral
temporal lobes and had at least one seizure recorded at 1 kHz sampling rate. The characteristics of ictal
HFO in the medial temporal lobe, and the relations between the presence of HFO, pathology, and postop-
erative seizure outcome were analyzed.
Results: Ictal HFO were detected from medial temporal structures in 11 patients with medial temporal
lobe epilepsy (MTLE). Among eight patients without HFO, only three were diagnosed with MTLE. Ictal
HFO were detected from unilateral medial temporal structures ipsilateral to the side of hippocampal scle-
rosis (HS). In one patient with bitemporal independent seizure onset, ictal HFO were detected only on the
side of HS. HS was detected in all 11 patients with HFO, but in only one of four patients without HFO.
Seizure outcome did not differ between patients with and without HFO.
Conclusions: Ictal HFO in the medial temporal lobe may be a specific marker for MTLE with HS.
Significance: Recording of ictal HFO in the medial temporal lobe may be useful for presurgical evaluation
of MTLE.
? 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights
reserved.
1. Introduction
Electrophysiological studies in animals revealed that high
frequency oscillations (HFO) with frequencies of 400–1000 Hz
were recorded in the vicinity of epileptogenic foci (Gastaut and
Fischer-Williams, 1959). Recently, HFO have attracted attention
in epilepsy surgery. Previous studies have reported the character-
istics of HFO in medial temporal lobe epilepsy (MTLE) as follows;
(1) recorded from hippocampus or entorhinal cortex; (2) frequen-
cies ranging from 170 to 400 Hz, usually with fast frequencies of
260–270 Hz [fast ripple: FR (Bragin et al., 1999), very high
frequency band: VHF (Jirsch et al., 2006)]; (3) usually detected
on the side of surgical resection; (4) not detected in the region of
secondary propagation; (5) can be recorded ictally or interictally;
(6) appear as clusters of short bursts with a duration of 6–53 ms;
(7) usually associated with ictal spikes on conventional EEG; (8)
detected by only 1–2 macroelectrode channels; and (9) low
amplitudes of 5–30 lV (Bragin et al.,1999, 2002; Jirsch et al.,
2006; Yamaguchi et al., 2008).
As for ictal HFO, Jirsch et al. (2006) reported ictal HFO in 10 pa-
tients with focal seizures, including four with medial temporal on-
set seizures. They used the EEG seizure onset as a surrogate for the
epileptogenic area, and did not evaluate the postoperative seizure
1388-2457/$36.00 ? 2011 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.clinph.2011.02.006
⇑Corresponding author. Address: National Epilepsy Center, Shizuoka Institute of
Epilepsy and Neurological Disorders, Urushiyama 886, Aoi-ku, Shizuoka 420-8688,
Japan. Tel.: +81 54 245 5446; fax: +81 54 247 9781.
E-mail address: n-usui@szec.hosp.go.jp (N. Usui).
Clinical Neurophysiology 122 (2011) 1693–1700
Contents lists available at ScienceDirect
Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph

Page 2

outcome. Khosravani et al. (2009) also studied ictal HFO in seven
TLE patients. However, the clinical significance of ictal HFO on sur-
gical decision-making has not been fully examined.
To clarify the clinical relevance of ictal HFO, we analyzed their
characteristics including the spatial distribution of ictal HFO in the
medialtemporallobe,andcomparedthepresenceorabsenceofictal
HFO with hippocampal pathology and postoperative seizure out-
come.Wealsodemonstratedtheclinicalusefulnessofdetectingictal
HFO in surgical decision-making for bitemporal epilepsy.
2. Patients and methods
2.1. Patients
Previously, we proposed the following criteria for omitting
intracranial EEG monitoring in patients with temporal lobe
epilepsy; (1) appearance of focal epileptic discharges in unilateral
sphenoidal lead during the phase of simple partial seizures, or uni-
lateral discharges predominantly in the sphenoidal lead during the
early phase of complex partial seizures; (2) interictal spikes on
scalp-recorded EEGs localizing unilaterally in the anterior region
of the temporal lobe, and if bilaterally independent, presenting
with unilateral predominance in a ratio of greater than 4:1; (3)
presence of autonomic signs in the initial phase of signal symp-
toms; and (4) neuroimaging findings in the mesial temporal region
showing elongated T2 on MRI and HS, or a tumorous lesion
(Mihara et al., 1992). Consequently, between May 2005 and May
2008, 19 patients underwent implantation of combined depth
and subdural electrodes in bilateral temporal lobes (Mihara and
Baba, 2001), and had at least one seizure recorded at 1 kHz sam-
pling rate. These 19 patients were included in this study. Before
intracranial video/EEG monitoring, all patients underwent non-
invasive presurgical evaluations including history-taking, video/
scalp sphenoidal EEG monitoring, neuroimagings, and neuropsy-
chological tests. Brain MRI was performed at 1.5 tesla and 5-mm
slice thickness, and axial, coronal, and sagittal T1-weighted,
T2-weighted, and FLAIR images were acquired. The clinical charac-
teristics of the 19 patients are shown in Table 1. Seizures were
classified by the semiological seizure classification proposed by
Lüders et al. (1998).
Depth electrodes (Unique Medical, Japan, 0.8 mm diameter,
1 mm length, either 5 or 10 mm center-to-center spacing) were
placed in bilateral hippocampi and amygdala using MRI stereotaxy.
Subdural electrodes (Ad-tech Medical Instrument, Racine, WI.
2.3 mm contact, effective area 4.15 mm2, 10 mm spacing) were
also placed over bilateral temporal regions including the basal
and anterior aspects, and the adjacent parieto-occipital areas
(Fig. 1). Reference electrodes were placed on the surface of the
skull, with the contacts of the electrodes facing away from the skull
to avoid the referential activation. Analyses were performed on ref-
erential montages. Antiepileptic medications were reduced, and
EEG recording was started approximately 1 week after electrode
placement and continued for approximately 2 weeks. The EEG sig-
nals were digitally recorded by EEG-1000 (Nihon Kohden) at a
sampling rate of 200 Hz and a time constant of 10 s for conven-
tional EEG analysis. For detection of HFO, EEG was low pass filtered
at 300 Hz, recorded at a sampling rate of 1 kHz and a time constant
of 10 s.
Table 1
Clinical characteristics of 19 patients.
PatientAge (yrs)/sexAge at onset (yrs) SemiologyInterictal EEG Ictal EEG
1
2
20/M
21/F
7
1
Automotor
Automotor
Lt
Bil
(Lt predominant)
Rt
Rt
Bil
(Lt predominant)
Bil
(Lt predominant)
Rt
Bil
(Lt predominant)
Rt
Bil
Bil
(Rt predominant)
Bil
Rt
Bil
Rt
Bil
Bil
Rt
Rt
3
4
5
32/F
36/F
43/F
10
15
9
Dysmnestic aura ? automotor
Aura ? automotor
Dialeptic
Rt
Lt
Lt
634/F1Aura ? automotorBil
(Lt predominant)
Lt
Bil
(Rt predominant)
Lt
Lt
Rt
7
8
30/F
19/M
4
3
Automotor
Abdominalaura ? automotor
9
10
11
43/F
37/F
39/M
9
18
10
Aura ? dialeptic
Automotor
Aura ? automotor
12
13
14
15
16
17
36/M
21/M
29/M
26/F
26/M
16/F
27
15
16
15
5
11
Dialeptic
Abdominal aura ? dialeptic
Aura ? automotor
Automotor
Hypermotor
Automotor
Lt
Rt
Lt
Rt
Bil
Bil
(Lt predominant)
Bil
(Lt predominant)
Lt
1831/F11Automotor Lt
1921/M 15Aura ? automotorBil
M: male; F: female; Rt: right; Lt: left; Bil: bilateral independent spikes on right and left.
Fig. 1. Skull X ray showing the location of intracranial electrodes. Left: anteropos-
terior view. Right: lateral view. Two depth electrodes are inserted into the medial
temporal structure on each side. RA and LA are aimed at amygdala, and RH and LH
are at hippocampus. Basal temporal areas are covered by subdural electrodes.
1694
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700

Page 3

Before the electrodes were removed, MRI was performed to
confirm the locations of the electrodes. MRI findings confirmed
that in all patients, the amygdalar depth electrodes were placed
correctly, and the hippocampal depth electrodes were placed in
the subiculum or Ammon’s horn.
After completing the invasive monitoring, 13 patients under-
went amygdalohippocampectomy, three anterior temporal lobec-
tomy, including medial temporal structures, and one lateral
temporal resection sparing the medial temporal structures. The
remaining two patients did not undergo resective surgery. Postop-
erative follow-up ranged from 12 to 54 months.
2.2. Visual inspection of ictal HFO
The ictal EEGs were analyzed visually for the presence of distinct
oscillations with frequencies of 200 Hz and higher (fast ripples) by
two clinical epileptologists (N.U. and K.T.). Both observers jointly
reviewed the data and established a consensus. High frequency
activitiesslowerthan200 Hz (ripples)werenot analyzedin thecur-
rent study. To visualize high frequency activities, the horizontal
(time) and vertical (amplitude) axes of the EEG display were ex-
panded, and the signals were digitally high-pass filtered at 50 Hz
(time constant of 0.003 s). HFO were defined as follows: (1) appear-
ing on several occasions at a similar frequency in the same channel;
(2) visually detectable as sinusoidal waves, and (3) containing at
least four consecutive peaks with similar inter-peak intervals. Typi-
cal examples of ictal HFO are shown in Figs. 2–4. The frequencies,
amplitudes, locations, durations and intervals of HFO were mea-
sured on the CRT screen. Moreover, total durations of HFO events,
aswellastheintervalsbetweenHFOonsetandEEGonsetonconven-
tional EEG, were also measured. Seizure onset on conventional EEG
wasdefinedaslocalized,sustained,rhythmic,orspikingEEGpattern
with a frequency >2 Hz, visually distinguished from background
activity (Spencer et al., 1992). Ictal EEG was reviewed 5 min before
EEG onset defined by conventional EEG. Interictal EEG of 5 min
was also reviewed in each patient. The state (awake or asleep) of
the patients during interictal recording was variable. In 15 patients,
theinterictaldataweremorethan2 hawayfromtheseizures.How-
ever, the interictaldata werewithin2 h (30 minaway fromseizures
in one patient, 40 min in two, and 1 h and 40 min in one) from sei-
zures in the remaining four patients. In patients who had interictal
HFO, the onset of ictal HFO were defined as the time at which HFO
appeared regularly and the interval of HFO became shorter than
3 s. In patients who had interictal HFO with a mean interval of less
than 3 s, the onset of ictal HFO were defined as the time at which
the mean interval of HFO became shorter than 1 s.
2.3. Correlation with hippocampal pathology and seizure outcome
In16patientswhounderwentmedialtemporalresection,thede-
greeofhippocampalneuronlossintheresectedspecimenswaseval-
uated. Hippocampal pathology was classified using Bl} umcke’s
classification (Bl} umcke et al., 2007) as follows: no MTS: normal hip-
pocampus, MTS type 1a: classic hippocampal sclerosis with severe
cell loss in CA1and moderate loss in remaining sectors, MTS type
1b: severe hippocampal sclerosis affecting all hippocampal sectors,
MTS type 2: severe cell loss in CA1 and only mild pathology within
remainingsectors(i.e.CA1-sclerosis),MTStype3:endfoliumsclero-
sis with moderate cell loss in all sectors with exception of CA1.
Seizure outcomes were evaluated using Engel’s criteria. The
relations between the presence of ictal HFO, pathology of resected
hippocampi, and postoperative seizure outcome were analyzed
statistically by using Fisher’s exact probability test. An error prob-
ability of less than 0.05 was considered to be indicative of
significance.
Table 2
Detection and parameters of ictal high frequency oscillations in 19 patients.
Patient
Seizure onset zone
Number of
seizures
Electrodes detecting
HFO
Frequency
(Hz)
Amplitude
(lV)
Duration
(ms)
Inter-HFO Interval (s)
Total duration
(s)
Time lag toEEG onset (s)
1
Lt MT (LA1-2, LH1-3, LBA1-2, LBP1-2)
3
LH1-3
200–250
25.4 (20.0–33.6)
75 (65–93)
1.15 (0.96–1.30)
30.3 (15–43)
23.2 (14–38)
2
Bil MT (RA1, RH1-2, RBP1-2, LA1-6, LH1-5, LBA1-6, LBP1-5)
3 (Rt onset)
RA1,RH1-2, RBP1
200–333
42.7 (17.8–73.2)
42 (32–61)
1.65 (1.50–1.96)
61.7 (26–91)
30.5 (8–53)
2 (Lt onset)
RA1,RH1-2, RBP1
200–333
39.4 (18.3–71.9)
40 (35–44)
0.70 (0.60–0.80)
79.5 (70–89)
74.5 (65–84)
3
Lt MT (LBA1, LBP1)
1
LH1-3
250–333
28.7 (12.5–44.7)
27.4 (23–38)
0.37 (0.25–0.64)
71
28
4
Rt MT (RA1-2, RH1-2, RBA1-3, RBP1-3)
11
RA1
200–333
35.3 (7.4–38.2)
19 (13–37)
0.82 (0.26–1.59)
28.8 (13–63)
20 (7–27)
5
Lt MT (LA1-2, LH1-3, LBP1-2)
2
LA1-2, LBP2
200–333
27.4 (14.7–44.5)
26 (24–28)
0.60 (0.56–0.65)
41.0 (30–52)
45.0 (38–52)
6
Lt MT (LA1-2, LH1-3, LBP1-2)
3
LH2-3, LBP1
200–333
22.8 (13.5–27.1)
32 (30–33)
1.11 (0.48–1.92)
45.0 (17–75)
33.7 (24–44)
7
Rt MT (RA1-2, RH1-2)
3
RA1-2, RH1-2,
200–333
21.6 (11.6–35.0)
27 (27–28)
1.13 (1.10–1.20)
68.3 (60–81)
28.5 (13.5–40)
LA1-2, LH1
200–250
17.1 (8.8–26.8)
30 (23–36)
0.34 (0.32–0.36)
0.5 (0.36–0.64)
-3.6 (3.5–3.6)
8
Lt MT (LA1-3, LH1-3, LBP1)
6
LH2, LBP1
200–250
17.3 (13.1–21.0)
56 (19–101)
1.47 (0.74–2.14)
65.7 (19–139)
21.0 (8–40)
9
Rt MT (RA1-2, RH2-3, RBA1, RBP1-2), Rt TP
5
RA1-2
200–333
35.8 (30.9–39.7)
31 (20–47)
0.8 (0.56–1.42)
55 (29–82)
11 (6–16)
10
Rt MT (RA1-2, RH1-3, RBP1-2)
3
RA1
200–250
26.1 (16.5–35)
32 (19–64)
0.84 (0.42–1.92)
48.7 (28–80)
13.3 (12–14)
11
Rt MT (RH1-2)
2
RH1
200–250
49.3 (11.8–89.7)
28 (25–38)
1.01 (0.4–2.18)
40.5 (38–43)
5 (0–10)
12
Lt MT (LA1-2, LH1-2)
2
–
–
–
–
–
–
–
13
Rt MT (RA1-2, RBP1-2)
1
–
–
–
–
–
–
–
14
Lt basal T (LH6, LBP5-6)
2
–
–
–
–
–
–
–
15
Rt MT (RH1, RBA1, RBP1)
1
–
–
–
–
–
–
–
16
Non-localizing
3
–
–
–
–
–
–
–
17
Lt T
2
–
–
–
–
–
–
–
18
Lt hemisphere
1
–
–
–
–
–
–
–
19
Lt basal T
2
–
–
–
–
–
–
–
Lt: left; Rt: right; Bil: bilateral; MT: medial temporal; TP: temporo-parietal; T: temporal.
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700
1695

Page 4

3. Results
3.1. Characteristics of ictal HFO
A total of 58 seizures were recorded at 1 kHz sampling rate.
The seizure onset zone identified on the conventional EEG record-
ing; number of seizures recorded at 1 kHz sampling rate; elec-
trodes detecting HFO; frequency, amplitude and duration of
HFO; inter-HFO interval; total duration; and time lag from HFO
onset to EEG seizure onset are shown in Table 2. Ictal HFO were
detected in 11 patients. In nine of 11 patients with HFO, all sei-
zures recorded at 1 kHz sampling originated from the medial
temporal regions ipsilateral to the side of surgical resection. In
Patient 2, independent seizure onset from the left and right med-
ial temporal regions was recorded. In Patient 9, five seizures were
recorded. Three seizures originated from the right medial tempo-
ral area, while the remaining two seizures arose from the right
temporo-parietal area preceding the discharges in the right med-
ial temporal area two to 3 days after withdrawal of carbamaze-
pine. Medial temporal onset seizures and temporo-parietal
onset seizures occurred on the same day. Finally, based on all
the information, including intracranial EEG findings, all 11 pa-
tients with ictal HFO in the medial temporal area were diagnosed
with MTLE and underwent resection surgery. Among the remain-
ing eight patients in whom no HFO were detected in the medial
temporal structures, only three were diagnosed with MTLE (pa-
tients 12, 13, and 15), while the seizure onset zones were heter-
ogeneous or poorly localized in the others. In one patient with no
ictal HFO recorded in the medial temporal lobe (Patient 19), ictal
HFO were detected in the basal temporal area (electrodes LBA5
and LBP5). Conventional ictal EEG also revealed seizure onset in
the left basal temporal region (LBA3-6, LBP3-6). Basal temporal
language area was also identified in the same region (LBA2-4,
LBP3). This patient did not undergo resection surgery, considering
the risk of language deficit.
The HFO in all 11 patients were segmental, lasted 19–75 ms
(mean value), were detected in 1–4 channels, and had frequencies
of 200–333 Hz. HFO could not be detected outside this frequency
range. In eight of 11 patients, HFO were detected at more than
one electrode. The mean amplitudes of HFO ranged from 17.3 to
49.3 lV. They were localized in unilateral medial temporal struc-
tures, mainly in the hippocampus (Fig. 2), and/or the amygdala.
In three patients (patients 4, 9, and 10), ictal HFO were exclusively
detected in the amygdala. HFO were also detected by subdural
electrodes placed over the mediobasal temporal region (parahip-
pocampal gyrus, TB) in five patients. HFO detected by subdural
electrodes always appeared simultaneously with the HFO detected
by depth electrodes. The electrodes with ipsilateral HFO were in-
cluded in the seizure onset zone defined by conventional EEG in
10 patients. In the remaining patient (Patient 3), the electrodes
with HFO were outside the seizure onset zone. However, both
the electrodes with HFO and seizure onset zone of the patient were
included in the left medial temporal structures and resected to-
gether by amygdalohippocampectomy.
HFO (ipsilateral to the side of surgical resection) onset preceded
conventional EEG ictal onset by 5.0–74.5 s (mean values). The
characteristics of HFO were consistent among seizures in each
patient.
Ictal HFO were usually associated with spikes. When HFO were
associated with spikes, they followed the peaks of the spikes. The
Fig. 2. Ictal EEG and HFO in Patient 1. (A) Conventional EEG (low-pass filter 120 Hz, time constant 0.1 s) reveals periodic sharp waves at LH1-3. The periodic sharp waves are
followed by more repetitive sharp waves in the left hippocampus and amygdala. Filled circles indicate the presence of HFO. (B) HFO waveforms (low-pass filter 300 Hz, time
constant 0.003 s). 200–250 Hz high frequency activities are detected at LH1-3. Note the calibration, which shows that HFO waves have very small amplitude and short
duration.
1696
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700

Page 5

mean total duration of ictal HFO ranged from 28.8 to 79.5 s. Con-
tralateral seizure propagation was seen in 10 of 11 patients with
ictal HFO. Except for one patient (Patient 7), HFO were not detected
in the areas of contralateral seizure propagation. In Patient 7, three
seizures were recorded at 1 kHz sampling rate. All seizures origi-
nated from the right medial temporal region, and propagated to
the left side. Ictal HFO were detected by the right medial temporal
electrodes (RA1-2, RH1-2). In this exceptional case, ictal HFO were
also detected in the left medial temporal electrodes (LA1-2, LH1) in
two of three seizures. HFO on the left side were observed only two
or three times, and the total durations of these contralateral HFO
were 0.36–0.64 s, much shorter than those of HFO recorded on
the resection side (Fig. 3).
In another patient (Patient 2), five seizures were recorded at
1 kHz sampling rate. On conventional EEG, three of five seizures
originated from the right medial temporal region, and the remain-
ing two from the left medial temporal region. Ictal HFO in the right
medial temporal structures were observed in the three seizures
originating on the right. Even in the two seizures originating from
the left side, ictal HFO were detected only in the right medial tem-
poral structures and preceded the left temporal EEG onset by 74.5 s
(Fig. 4).
Interictal HFO were also detected in 10 patients (Patients 2–5,
7–11, and 18) (Table 3). These interictal HFO appeared more
irregularly, and the inter-burst intervals of HFO lasted longer than
those of ictal HFO. In eight patients, interictal HFO were detected
only on the side of resection. In one patient (Patient 11), interictal
HFO were detected in bilateral hippocampi, although the fre-
quency of appearance was much higher on the ipsilateral side.
In the remaining patient who did not proceed to surgery (Patient
18), interictal HFO were detected in the unilateral medial tempo-
ral structure.
3.2. Ictal HFO and hippocampal pathology
Surgical procedures and hippocampal pathology are shown in
Table 4. In 11 patients who had ictal HFO, histological examination
revealed HS in all the patients (type 1a in nine, and type 2 in two).
In three patients (Patients 4, 9, and 10) showing ictal HFO exclu-
sively in the amygdala, HS (type 1a) was also identified histologi-
cally. In eight patients who showed no ictal HFO, five underwent
medial temporal resection (amygdalohippocampectomy in three,
anterior temporal lobectomy including medial temporal structures
in two), and pathological results were available in four of five pa-
tients. Classic sclerosis (type 1a) was observed in one patient,
while no HS was found in the remaining three patients. The asso-
ciation between the presence of ictal HFO and HS was statistically
significant (p = 0.008791).
3.3. Ictal HFO and postoperative seizure outcome
Favorable seizure outcome (Engel’s class I) was obtained in
eight of 11 patients with ictal HFO (Table 4). In one patient with
class II outcome (Patient 11), a few seizures were observed soon
Fig. 3. Ictal EEG and HFO in Patient 7. (A) On conventional EEG (low-pass filter 120 Hz, time constant 0.1 s), ictal EEG starts with a large positive sharp wave at RH1 and RH2,
followed by beta activities at the same electrodes. Seven seconds later, rhythmic beta waves start at RA1 and RA2 with gradual evolution. Approximately 45 s later, left-sided
electrodes (LH1-4 and LA1-4) start showing small spikes, followed by beta activities in the same electrodes Filled circles and the bold line indicate the presence of HFO. (B)
HFO waveforms before conventional EEG onset, around onset, and around conventional EEG onset at the left hemisphere (low-pass filter 300 Hz, time constant 0.003 s). These
demonstrate 200–333 Hz high-frequency activities at 1–4 electrodes in the right side (RA1, RA2, RH1 and RH2). These also demonstrate high-frequency activities at LA1-2 and
LH1 around the conventional EEG onset on the left side. Circles indicate the locations of HFO.
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700
1697

Page 6

after surgery, and the seizures were controlled by increasing anti-
epileptic medications. In one patient (Patient 9) with poor outcome
(class III), seizures arising from contralateral temporal region were
documented postoperatively. The reason for unfavorable outcome
in the remaining patient (Patient 3) was unclear. Among the three
cases with ictal HFO restricted to amygdala (Patients 4, 9, and 10),
favorable outcome was obtained in two, and unfavorable outcome
in one. Favorable seizure outcome (Engel’s class I) was obtained in
three of five patients without HFO who underwent resective sur-
gery including medial temporal structures. Comparing the patients
who underwent amygdalohippocampectomy in the two groups,
favorable seizure outcome was obtained in seven of 10 patients
with ictal HFO, and in two of three patients without HFO. Seizure
outcome was not statistically different between the patients with
and without ictal HFO.
4. Discussion
4.1. Characteristics of ictal HFO in medial temporal lobe
HFO in medial temporal structures have been reported in recent
years. Most reports of HFO in patients with MTLE analyzed the
interictal state with microelectrodes, usually during non-REM
sleep (Staba et al., 2004). Fast ripples are considered to represent
hypersynchronous discharges of locally interconnected principle
neurons capable of generating spontaneous seizures (Staba et al.,
2002). Fast ripples are most often recorded on a single microwire,
supporting the hypothesis that fast ripple HFO are primarily gener-
ated by highly localized, sub-millimeter scale neuronal assembles
that are most effectively sampled by microwire electrodes (Worrell
et al., 2008). However, other reports have demonstrated that mac-
roelectrodes also detect HFO in medial temporal structures (Jirsch
et al., 2006; Khosravani et al., 2009; Yamaguchi et al., 2008).
Although interictal HFO in the medial temporal lobe have been
widely studied, the clinical significance of ‘ictal’ HFO in the medial
temporal lobe has not been specifically studied.
In this study, multiple sites of the medial temporal regions were
investigated with both depth and subdural electrodes. Ictal HFO
appeared in a relatively stereotyped segmental fashion, and were
detected mainly in the hippocampus and/or the amygdala. They
were mostly located unilaterally ipsilateral to the side of HS, in
contrast to interictal HFO in medial temporal lobe, which can be
detected bilaterally (Staba et al., 2002). In eight patients, ictal
HFO were detected at multiple electrode contacts, suggesting that
HFO may reflect synchronous discharges widely located within
unilateral medial temporal region. The sensitivity of subdural elec-
trodes in detecting HFO was lower than that of depth electrodes.
Fig. 4. Ictal EEG and HFO in Patient 2. (A) On conventional EEG (low-pass filter
120 Hz, time constant 0.1 s), ictal EEG starts with repetitive spikes at LA1, LH2, LH3,
and LH4, as well as rhythmic alpha activity at LH1 at the same time. The spikes
spread widely in the left temporal area, and are replaced by rhythmic beta waves at
LH1, LH2, LH3, and LH4. The right hemispheric electrodes do not show clear ictal
activities throughout. Filled circles and bold line indicate the presence of HFO. (B)
HFO waveforms before conventional EEG onset, and around onset (low-pass filter
300 Hz, time constant 0.003 s). 200–333 Hz high frequency activities are recorded
at 1 to 3 electrodes on the right side (RA1, RH1, and RH2). Circles indicate the
locations of HFO. Although spikes are seen on the left side, HFO is not recorded at
the left-sided electrodes.
Table 3
Detection and parameters of interictal high frequency oscillations.
Patient Electrodes detecting HFO Frequency (Hz)Amplitude (lV)Duration (ms) Inter-HFO Interval (s)
1
2
3
4
5
6
7
8
9
10
11
None
RH1-2
LH2-3
RA1
LA1-2, LBP1-2
None
RA1
LH2, LBP1
RA1-2
RA1
RH1
LH2
None
None
None
None
None
None
LH1–2, LBP1–2
None
–
200–333
200–333
200–333
200–333
–
200–333
200–250
200–333
200–250
200–250
200–250
–
–
–
–
–
–
200–333
–
–
48.5 (30.9–70.6)
42.7 (17.6–61.8)
25.0 (10.3–88.2)
39.5 (22.6–55.8)
–
42.7 (20.6–85.3)
29.4 (26.5–35.3)
32.0 (12.0–39.7)
29.4 (20.6–38.2)
57.8 (44.1–76.5)
35.3 (26.5–44.1)
–
–
–
–
–
–
26.8 (13.2–41.2)
–
–
31 (33–71)
25 (19–28)
30 (18–43)
23 (16–32)
–
22 (13–34)
50 (29–67)
25 (19–34)
19 (15–25)
26 (10–46)
20 (17–23)
–
–
–
–
–
–
27 (22–32)
–
–
24.4 (6.4–42.9)
1.5 (0.3–3.3)
12.1 (1.4–73)
2.5 (0.4–17.9)
–
18.3 (10.6–23.7)
4.3 (1.6–18.7)
2.41 (0.4–6.8)
27.8 (3.6–121)
4.3 (0.8–16.3)
90 (40–199)
–
–
–
–
–
–
31 (1–131)
–
12
13
14
15
16
17
18
19
1698
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700

Page 7

Depth electrodes were inserted into subiculum, Ammon’s horn,
and amygdala, whereas subdural electrodes were implanted over
the parahippocampal gyrus. One reason for the different sensitivity
may be the different locations of the electrodes. Subiculum, Am-
mon’s horn, and amygdala may generate ictal HFO more frequently
than parahippocampal areas. Another reason may be the wider
contact surface of subdural electrode (4.15 mm2) than that of
depth electrode (2.6 mm2, calculated from the diameter and the
length). In addition to the location of electrodes, the difference in
contact surface of the electrodes may affect the sensitivity.
In 10 of 11 patients with ictal HFO, the electrodes with ipsilat-
eral ictal HFO were included in the seizure onset zone, whereas in
the remaining patient, they were outside the seizure onset zone.
Since both the areas with HFO and seizure onset zone were re-
sected, whether HFO or seizure onset zone is superior for deciding
the area of resection cannot be determined.
Usually,HFOfollowedthepeaksofthespikes.Inaseparatestudy
of neocortical epilepsy, we found that very high frequency activities
faster than 1000 Hz usually preceded the spikes. In other words,
spikesseemtointerruptveryhighfrequencyactivitiesinneocortical
epilepsy. Therefore, we speculate that spikes may be an inhibitory
phenomenon in neocortical epilepsy (Usui et al., 2010). On the con-
trary, the temporal relation between spikes and HFO in the medial
temporal lobe suggests that the spikes may trigger HFO. Therefore,
spikes may be an excitatory phenomenon in MTLE.
In a conventional EEG setting, inter-observer differences are
common even for the interpretation of ictal onset, since conven-
tional EEG waveforms can be very variable. On the contrary, the
characteristics of medial temporal ictal HFO were very consistent
among patients. Therefore, the identification of ictal HFO may sup-
port interpretation of ictal EEG.
Ictal HFO (ipsilateral to the side of surgical resection) always
appeared before conventional EEG seizure onset. Except in one pa-
tient, ictal HFO did not appear in the region of contralateral spread.
Moreover, they did not spread outside the medial temporal re-
gions. These findings suggest that HFO are strongly related to epi-
leptogenicity. Jirsch et al. (2006) also reported concordant results.
In one exceptional case, HFO appeared in the area of contralateral
seizure spread, but the total duration of contralateral HFO was
much shorter. In this case, amygdalohippocampectomy on the side
of seizure origin was performed, and the patient has been seizure
free for approximately 2 years. Careful follow-up of this patient is
necessary to evaluate epileptogenicity in the contralateral medial
temporal lobe.
Temporal lobe epilepsy has been considered as a bilateral dis-
ease (Margerison and Corsellis, 1966). It is often difficult to decide
the surgical indication and the side of resection in patients with
bitemporal epilepsy, even with intracranial EEG. In one patient,
clinical seizures originated independently from bilateral medial
temporal lobes when assessed on conventional EEG. However, ictal
HFO were detected unilaterally. The findings in this patient
strongly suggest that ictal HFO may not be just a part of ictal
EEG changes, but probably represent electrophysiological phenom-
ena independent of the activity usually detected by conventional
EEG, and that epileptogenicity and ictogenicity could be differenti-
ated by ictal HFO. Crépon et al. (2010) recently studied interictal
HFO with macroelectrodes, and reported detection of interictal
HFO unilaterally in a case of bitemporal seizure onset. They sug-
gested the usefulness of interictal HFO for defining the seizure on-
set zone. In our patients, interictal HFO were detected in 10
patients, and they were unilateral and ipsilateral to the side of
resection in eight patients. Therefore, interictal HFO may be also
useful for deciding the side of resection. However, interictal HFO
in medial temporal lobe has also been detected bilaterally (Staba
et al., 2002). The one reason for disagreement may be the extended
periods of recording in the study by Staba et al. (2002) (the mean
length of time analyzed was 177 ± 7 min of non-REM sleep) com-
pared with our study (5 min). Further studies are necessary for
clarifying the usefulness of ictal and interictal HFO for deciding
the side of resection in bitemporal epilepsy.
4.2. Correlation with hippocampal pathology
Using microelectrodes, Staba et al. (2007) reported that higher
fast ripples to ripple ratios are associated with histopathologic
changes found in HS in TLE patients. Recently, Ogren et al. (2009)
studied interictal HFO recorded by hippocampal microelectrodes
in10patients withMTLE,anddemonstrated theproximitybetween
fastripplesandlocalhippocampalatrophy.Ourstudydemonstrated
a strong association between ictal HFO (fast ripple range) in medial
temporal lobe detected by macroelectrodes and HS. In Patient 11,
preoperative MRI did not reveal clear HS. However, ictal HFO were
detected in the right hippocampal electrode and HS was detected
pathologically. Ictal HFO in the medial temporal lobe may be the
electrophysiological signature of HS. Neuron loss and synaptic reor-
ganization may contribute to the generation of HFO (Ogren et al.,
2009).
Table 4
Surgery, hippocampal pathology, and seizure outcome in 19 patients.
Patient Surgery Follow-up (months)MRI
Pathology (Bl} umcke)
Outcome (Engel)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Lt AHE
Rt AHE
Lt AHE
Rt AHE
Lt ATL
Lt AHE
Rt AHE
Lt AHE
Rt AHE
Rt AHE
Rt AHE
Lt AHE
Rt AHE
Lt LTR
Rt AHE
Lt ATL
Lt ATL
No
No
49
51
13
43
33
12
22
26
24
12
17
54
35
37
48
40
24
–
–
Lt HS
Rt HS
Lt HS
Rt HS
Lt HS
Lt HS
Rt HS
Lt HS
Rt HS
Rt HS
Normal
Normal
Rt AH
Normal
Rt HS
Normal
Normal
Lt HS
Normal
HS (Type 1a)
HS (Type 1a)
HS (Type 1a)
HS (Type 1a)
HS (Type 1a)
HS (Type 2)
HS (Type 2)
HS (Type 1a)
HS (Type 1a)
HS (Type 1a)
HS (Type 1a)
No HS
No HS
NA
HS (Type 1a)
No HS
NA
–
–
Ia
Id
IIIa
Ic
Id
Ia
Ia
Ib
IIIa
Ia
IIb
IIb
Ia
IIIa
Ic
Ia
IIIa
–
–
Rt: right; Lt: left; AHE: amygdalohippocampectomy; ATL: anterior temporal lobectomy; LTR: lateral temporal resection; HS: hippocampal sclerosis; AH: amygdalar
hypertrophy; NA: data not available.
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700
1699

Page 8

Among the four patients without ictal HFO who underwent
resection of the medial temporal lobe, histopathology revealed
no HS in three of four patients. In Patient 13 with no HS, amygdalo-
hippocampectomy was performed and the patient became seizure
free. Ictal HFO may be absent in patients with MTLE without HS. In
one patient who had HS (classic sclerosis) but no HFO (Patient 15),
seizures continued with somatomotor signs for 2 years after amy-
gdalohippocampectomy, and then became controlled without
medication adjustment, and she has been seizure-free for more
than 2 years (Engel’s class Ic). Although the reason for late seizure
remission in this patient is not clear, the epileptogenic zone might
not be restricted to medial temporal structures. In addition to HS,
primary epileptogenicity in the medial temporal lobe may be a req-
uisite for generating ictal HFO.
4.3. Correlation with postoperative seizure outcome
Eight of 11 patients with HFO achieved Engel’s class I outcome,
and one patient had class II outcome. In one case (Patient 9), con-
tralateral temporal onset seizures were recorded after surgery,
although contralateral HFO was never recorded during invasive
monitoring. Although an association between resection of ictal
HFO generating zone and favorable seizure outcome in neocortical
epilepsy has been suggested (Ochi et al., 2007), such a relationship
could not be proven in this study. However, this result should be
interpreted with caution. A strong association with ictal HFO and
HS was established in this study. All patients with ictal HFO were
diagnosed with MTLE. Subsequently, seven of 10 patients with
HFO who underwent amygdalohippocampectomy obtained good
seizure outcome. On the contrary, only three of eight patients
without HFO were diagnosed with MTLE. In other patients, the epi-
leptogenic zone was heterogeneous or poorly localized. The pres-
ence of ictal HFO in the medial temporal lobe strongly favors a
diagnosis of MTLE with HS and subsequent resective surgery. The
small number of patients without HFO may lead to the negative re-
sult for seizure outcome. Further study including more patients
without HFO may clarify the relationship between the presence
of ictal HFO and postoperative seizure outcome.
Jacobs et al. (2010) analyzed interictal HFO in 20 patients, and
compared rates and extents of HFO in resected and non-resected
areas with surgical outcome. Patients with a good outcome (Class
I or II) had a significantly larger proportion of HFO-generating
areas removed than patients with a poor outcome (Class III or
IV). In our 11 patients with ictal HFO, HFO-generating areas were
completely resected in all 11 patients. However, there were two
patients with a poor outcome (Class III). The discordance may be
due to the fact that Jacobs et al. included both fast ripples and rip-
ples as HFO, whereas we did not include ripples as HFO. Ripples
can be either pathologic or memory-related. Further studies for dif-
ferentiating pathological ripples from healthy, memory-related
ones are necessary.
4.4. Clinical significance of absence of HFO in the medial temporal lobe
In eight patients, no ictal HFO was detected in the medial tem-
poral lobe. Compared to patients with HFO, the epileptogenic
zones of patients without HFO were more heterogeneous or poorly
localized.
When no ictal HFO is detected in the medial temporal area, two
possibilities should be considered. One is that the diagnosis is not
MTLE. In Patient 16, lateral temporal lobe epilepsy was diagnosed
and left anterior temporal lobectomy, including medial temporal
lobe,wasperformedwithcompleteseizurecontrol,andhistopathol-
ogy revealed no HS. In this case, absence of ictal HFO in the medial
temporal lobe may simply reflect that the medial temporal lobe
might not be included in the epileptogenic zone. In Patient 19 with
no HFO in the medial temporal lobe, ictal HFO were detected in the
basal temporal lobe. Ictal high frequency activities recorded from
neocortex have been reported (Jirsch et al., 2006; Ochi et al., 2007;
Usui et al., 2010). However, most of the patients in this study were
diagnosedwithpossibletemporallobeepilepsy.Thereforeneocorti-
calHFOwasnotcommon.Anotherpossibilityisthatthediagnosisis
MTLE without HS. In Patient 13, amygdalohippocampectomy re-
sulted in seizure control, and histopathology revealed no HS. There-
fore, in some patients with MTLE without HS, ictal HFO may not be
detected.Althoughfurtherresearchisneeded,ictalHFOinthemed-
ial temporal lobe may be a specific marker for MTLE with HS.
References
Bl} umcke I, Pauli E, Clusmann H, Schramm J, Becker A, Elger C, Merschhemke M,
Meencke H, Lehmann T, Von Deimling A, Scheiwe C, Zentner J, Volk B, Romst} ock
J, Stefan H, Hildebrandt M. A new clinico-pathological classification system for
mesial temporal sclerosis. Acta Neuropathol. 2007;113:235–44.
Bragin A, Engel J, Wilson CL, Fried I, Buzsáki G. High-frequency oscillations in
human brain. Hippocampus 1999;9:137–42.
Bragin A, Wilson CL, Staba RJ, Reddick M, Fried I, Engel Jr J. Interictal high-frequency
oscillations (80–500 Hz) in the human epileptic brain: entorhinal cortex. Ann.
Neurol. 2002;52:407–15.
Crépon B, Navarro V, Hasboun D, Clemenceau S, Martinerie J, Baulac M, Adam C, Le
Van Quyen M. Mapping interictal oscillations greater than 200 Hz recorded
with intracranial macroelectrodes in human epilepsy. Brain 2010;133:33–45.
Gastaut H., Fischer-Williams, M. The physiopathology of epileptic seizures.
American PhysiologicalSociety, Handbook
comprehensive presentationof physiological
American Physiological Society, Washington DC, 1959, pp. 329–363.
Jacobs J, Zijlmans M, Zelmann R, Chatillon C-E, Hall J, Olivier A, Dubeau F, Gotman J.
High-frequency electroencephalographic oscillations correlate with outcome of
epilepsy surgery. Ann. Neurol. 2010;67:209–20.
Jirsch JD, Urrestarazu E, Levan P, Olivier A, Dubeau F, Gotman J. High-frequency
oscillations during human focal seizures. Brain 2006;129:1593–608.
Khosravani H, Mehrotra N, Rigby M, Hader W, Pinnegar CR, Pillay N, Wiebe S,
Federico P. Spatial localization and time-dependent changes of electrographic
high frequency oscillations in human temporal lobe epilepsy. Epilepsia
2009;50:605–16.
Lüders H, Acharya J, Baumgartner C, Benbadis S, Bleasel A, Burgess R, Dinner DS,
Ebner A, Foldvary N, Geller E, Hamer H, Holthausen H, Kotagal P, Morris H,
Meencke HJ, Noachtar S, Rosenow F, Sakamoto A, Steinhoff BJ, Tuxhorn I, Wyllie
E. Semiological seizure classification. Epilepsia 1998;39:1006–13.
MargerisonJH, CorsellisJAN.Epilepsy
1966;89:499–530.
Mihara T, Baba K. Combined use of subdural and depth electrodes. In: L} uders HO,
Comair YG, editors. Epilepsy surgery. second ed. Philadelphia: Lippincott
Williams & Wilkins; 2001. p. 613–21.
Mihara T, Matsuda K, Tottori T, Baba K, Inoue Y, Hiyoshi T, Watanabe Y, Yagi K, Seino
M. Conditions for omitting intracranial long-term monitoring before surgical
resection in patients with temporal lobe epilepsy. Jpn. J. Psychiatry Neurol.
1992;46:323–9.
Ochi A, Otsubo H, Donner EJ, Elliott I, Iwata R, Funaki T, Akizuki Y, Akiyama T, Imai K,
Rutka JT, Snead III OC. Dynamic changes of ictal high-frequency oscillations in
neocortical epilepsy: using multiple band frequency analysis. Epilepsia
2007;48:286–96.
Ogren JA, Wilson CL, Bragin A, Lin JJ, Salamon N, Dutton RA, Luders E, Fields TA, Fried
I, Toga AW, Thompson PM, Engel Jr J, Staba RJ. Three-dimensional surface maps
link local atrophy and fast ripples in human epileptic hippocampus. Ann.
Neurol. 2009;66:783–91.
Spencer SS, Guimaraes P, Katz A, Kim J, Spencer D. Morphological patterns of
seizures recorded intracranially. Epilepsia 1992;33:537–45.
Staba RJ, Wilson CL, Bragin A, Fried I, Engel Jr J. Quantitative analysis of high-
frequency oscillations (80–250 Hz) recorded in human epileptic hippocampus
and entorhinal cortex. J. Neurophysiol. 2002;88:1743–52.
Staba RJ, Wilson CL, Bragin A, Jhung D, Fried I, Engel Jr J. High-frequency oscillations
recorded in human medial temporal lobe during sleep. Ann. Neurol.
2004;56:108–15.
Staba RJ, Frighetto L, Behnke E, Mathern GW, Fields T, Bragin A, Ogren J, Fried I,
Wilson CL, Engel Jr J. Increased fast ripple to ripple ratios correlate with reduced
hippocampal volumes and neuron loss in temporal lobe epilepsy patients.
Epilepsia 2007;48:2130–8.
Usui N, Terada K, Baba K, Matsuda K, Nakamura F, Usui K, Tottori T, Umeoka S,
Fujitani S, Mihara T, Inoue Y. Very high frequency oscillations (over 1000 Hz) in
human epilepsy. Clin. Neurophysiol. 2010;121:1825–31.
Worrell GA, Gardner AB, Stead SM, Hu S, Goerss S, Cascino GJ, Meyer FB, Marsh R,
Litt B. High-frequency oscillations in human temporal lobe: simultaneous
microwire and clinical macroelectrode recordings. Brain 2008;131:928–37.
Yamaguchi M, Terada K, Baba K, Nakamura F, Inoue Y. Ictal high frequency
oscillations in patients with mesial temporal lobe epilepsy. Jpn. J. Clin.
Neurophysiol. 2008;36:615–23.
of Physiology:
knowledge
A critical,
concepts.and
in the temporallobes.Brain
1700
N. Usui et al./Clinical Neurophysiology 122 (2011) 1693–1700