Intrinsic Epileptogenicity of

Human Dysplastic Cortex as Suggested by

Corticography and Surgical Results
Andre Palmini, MD," Antonio Gambardella, MD,I-§ Frederick Andermann, MD, FRCP(C),t
Francois Dubeau, MD, FRCP(C),i Jaderson C. da Costa, MD, PhD," Andre Olivier, MD, PhD, FRCS(C),t
Donatella Tampieri, MD,? Pierre Gloor, MD, PhD,I Felipe Quesney, MD, PhD,? Eva Andermann, MD, PhD,?
Eduardo Paglioli, MD," Eliseu Paglioli-Neto, MD," Ligia Coutinho, MD, PhD,"
kchard Leblanc, MD, FRCS(C),t and Hyoung-Ihl Kim, MD, PhDS

Cortical dysplastic lesions (CDyLs) are often associated with severe partial epilepsies. We describe the electrographic
counterpart of this high degree of epileptogenicity, manifested by continuous or frequent rhythmic epileptogenic
discharges recorded directly from CDyLs during intraoperative electrocorticography (ECoG).These ictal or continuous
epileptogenic discharges (I/CEDs) assumed one of the following three patterns: (I) repetitive electrographic seizures,
(2) repetitive bursting discharges, or ( 3 )continuous or quasicontinuous rhythmic spiking. One or more of these patterns
were present in 23 of 34 patients (67%) with intractable partial epilepsy associated with CDyLs, and in only 1 of 40
patients (2.5%) with intractable partial epilepsy associated with other types of structural lesions. I/CEDs were usually
spatially restricted, thus contrasting with the more widespread interictal ECoG epileptic activity, and tended to
colocalize with the magnetic resonance imaging-defined lesion. Completeness of excision of cortical tissue displaying
IiCEDs correlated positively with surgical outcome in patients with medically intractable seizures; i.e., three-fourths
of the patients in whom it was entirely excised had favorable surgical outcome; in contrast, uniformly poor outcome
was observed in those patients in whom areas containing I/CEDs remained in situ. We conclude that CDyLs are highly
and intrinsically epileptogenic, and that intraoperative ECoG identification of this intrinsically epileptogenic dysplastic
cortical tissue is crucial to decide the extent of excision for best seizure control.
Palmini A, Gambardella A, Andermann F, Dubeau F, da Costa JC, Olivier A, Tampieri D, Gloor P, Quesney F,
Andermann E, Paglioli E, Paglioli-Net0 E, Coutinho L, Leblanc R, Kim H-I. Intrinsic epileptogenicity of human
dysplastic cortex as suggested by corticography and surgical results. Ann Neurol 1995;37:476-487

Intractable partial epilepsy is often associated with
structural lesions of the cerebral cortex [ 1-51, Among
the different types of structural epileptogenic abnor-
malities, cortical dysplastic lesions (CDyLs) have gener-
ated a great deal of interest lately, since many can now
be reliably detected during life by magnetic resonance
imaging (MRI) [6- 101. High-resolution MRI has
shown that areas of focally increased cortical thickness,
reduced gyration, and lack of gray-white matter digita-
tion are more frequent than previously suspected and
are responsible for many epileptic disorders previously
considered cryptogenic or idiopathic [ 11, 1.21. Micro-
scopically, CDyLs are characterized by a lack of normal
cortical lamination, which may be associated with ab-
normal giant neurons, and occasionally also with bi-
zarre large eosinophilic cells, so-called balloon cells
113- 151. Dendrites and axons are often abnormally
oriented and distributed [ 161. The epileptogenic po-
tential of these abnormalities of cortical architecture is
only now beginning to be elucidated (151.
Recently, we observed during acute intraoperative
electrocorticography (ECoG) that some patients with
CDyLs undergoing epilepsy surgery displayed pro-
longed trains of rhythmic epileptogenic activity of
various patterns [17]. These ictal or continuous epi-
leptogenic discharges (I/CEDs) were spatially more
restricted than the more diffuse interictal spiking. They
were recorded from ECoG electrodes overlying both
the visible dysplastic cortex and normal appearing ad-
jacent neocortex. This observation was unexpected,

From the *Port0 Alegre Epilepsy Surgery Program, Neurology and
Neurosurgery Services, Hospital Sao Lucas da PUCRS, Porto Ale-
gre, Brazil; +Department of Neurology and Neurosurgery, McGill
University, and the Montreal Neurological Institute and HospitaL
Montreal, Ca.iada; and $Department of Neurosurgery, Chonbuk
National Uigiversity Hospital, Chonju, Korea.
Received JUI 8, 1994, and in revised form Nov 10. Accepted for
publication Nov 10, 1994.
Address correspondence to Dr Palmini, Servico de Neurologia, Hos-
pital Sao Lucas-PUCRS, Av Ipiranga 6690, Porto Alegre RS, Brasil
CEP 906 10-000.
gPresent address:Isrituto de scienze
Neuro~ogice, 88100 Catanzaro,
Italy.
This paper was awarded the Cesare Lomhroso Prize of the Brazilian
League Of Epilepsy, Campinas,szo Paula, July 1994,

476 Copyright 0 1995 by the American Neurological Association

since structural abnormalities associated with chronic
neocortical epilepsy are usually considered to be epi-
leptogenic due to their effect upon adjacent cortex.
They are not epileptogenic in themselves { 181. The
effects of structural lesions such as tumors, cysts, or
arteriovenous malformations on adjacent neocortex are
not well understood, and a number of putative mecha-
nisms including pressure effects, localized ischemia,
and neurochemical changes have been invoked to ex-
plain epileptogenicity E4, 5 , 191. It is, however, gener-
ally accepted that these lesions are “epileptogenic” be-
cause they interfere with the normal physiology of the
adjacent neocortex. The recording of I/CEDs directly
from dysplastic cortex itself suggested that, in contrast
with most other structural epileptogenic neocortical/
lesions, CDyLs might have intrinsic epileptogenicity.
We conducted the present study in a series of pa-
tients with CDyLs and intractable partial epilepsy, with
the following objectives: (1) to further describe the
various patterns of I/CEDs associated with CDyLs; (2)
to examine the frequency of occurrence of this finding
in CDyLs in comparison with other structural epilepto-
genic lesions; and (3) to evaluate the role of I/CEDs
in the planning of surgical resection for patients with
CDyLs and intractable epilepsy. This latter aim was
of particular interest to us, since we have previously
reported that the extent of removal of the more wide-
spread interictal spiking on ECoG did not correlate
with surgical outcome in these patients, whereas the
best results were obtained when most or all of the
visible lesion could be resected C201.
Patients and Methods
Two groups of patients were studied. The cortical dysplasia
group (CDyG) originally consisted of 34 patients with local-
ized CDyLs and intractable partial epilepsy, consecutively
evaluated and operated for seizure control at three different
centers, as follows: 27 patients at the Montreal Neurological
Institute (MNI), from 1975 to 1991; 6 at the Epilepsy Sur-
gery Program, in Porto-Alegre, Brazil, from April 1991 to
April 1994; and 1 from the Chonbuk National University
Hospital, in Chonju, Korea, in 1993. Nine additional pa-
tients were excluded since intraoperative ECoG recordings
were either not performed, did not cover the dysplastic le-
sion, o r were not available for review. All patients of this
group had unequivocal imaging or histological diagnosis of a
localized CDyL [6, 211. In 2 patients these abnormalities
involved one hemisphere diffusely and constituted examples
of hemimegalencephaly. Microscopically, CDyLs were char-
acterized by cortical dyslamination, giant “dysplastic” neu-
rons, and, occasionally, very large cells, with eosinophilic cy-
toplasm and immunostaining profile intermediate between
bizarre astrocytes and neurons (“balloon cells”) [l 3, 141.
Such cells are commonly found in patients with tuberous
sclerosis {22), and when present in the context of a single,
localized CDyL, have led to the use of the term “forme fruste
of tuberous sclerosis” [9,23,241. The presence of these cells
suggests a more severe degree of histological abnormality in
Palmini et al: Epileptogenicity of Dysplastic Cortex
CDyLs and should be viewed in this context [ 2 5 ) . In addi-
tion, most patients had a variable number of dysplastic neu-
rons scattered through the subcortical white matter. Seven-
teen of these 34 patients were included in previous reports
[9, 20). Sixteen were male and 18 female, with ages at the
time of operation ranging from less than 1 to 38 years (mean,
16.5 yr; SD, 9.2 yr). Age at onset of epilepsy ranged from
less than 1 month to 19 years (mean, 5.0 yr; SD, 5.1 yr), and
the duration of epilepsy prior to surgery ranged from 3
months to 36 years (mean, 11.7 yr; SD, 8.1 yr).
The second group consisted of 40 patients who were evalu-
ated for intractable partial epilepsy and were found to harbor
nondysplastic structural lesions. Thirty were studied at the
MNI, randomly selected, and all had a cortical tumor oper-
ated under ECoG guidance in the last 5 years; 3 had oligo-
dendrogliomas, 4 mixed gliomas, and the remaining 2 3 grade
I to 111 astrocytomas. The other 10 patients with structural
nond ysplastic epileptogenic lesions were consecutively evalu-
ated and operated at the Porto Alegre Epilepsy Surgery Pro-
gram, from February 1992 to April 1993. Six had cortical
tumors ( 5 grade 1-11 astrocytomas, 1 oligodendroglioma),
and 1 each had a nonprogressive calcified lesion, a chronic
inflammatory cyst, a postmeningitic atrophic scar, and a pro-
gressive atrophic lesion associated with chronic inflammatory
changes, suggestive of Rasmussen’s encephalitis. This non-
dysplasia group (NonDyG) consisted of 21 men and 19
women, with ages at operation and at onset of epilepsy rang-
ing respectively from 6 to 40 years (mean, 24.8 yr; SD, 10.1
yr) and 6 months to 39 years (mean, 14.5 yr; SD, 13.1 yr).
Epilepsy duration ranged from 1 to 24 years (mean, 11.2 yr;
SD, 6.04 yr).
All patients from both groups had 16-channel extracranial
electroencephalograms (EEGs) with the international 10-20
electrode system, recorded directly or by cable telemetry.
Sphenoidal and supraorbital electrodes were used when indi-
cated and prolonged recordings with video monitoring were
always performed during wakefulness and sleep. The pres-
ence, distribution, and pattern of interictal spikes and sharp
waves were noted. In particular, we recorded the presence
of rhythmic trains of repetitive spikes, at variable frequencies.
Two patients from the CDyG also had preoperative chronic
recordings with 64-contact subdural grids. Neuropsychologi-
cal studies [26] were performed in all patients except in 4 of
the CDyG and 1 in the NonDyG who were less than 6 years
old at the time of evaluation. Intracarotid amobarbitai sodium
testing was carried out in 24 patients of the CDyG and in
39 of the NonDyG, according to established protocols [27).
Computed tomographic (CT) scans were available in all
patients. Twenty-six patients of the CDyG had MRIs (2 with
a 0.5-T Philips Gyroscan, 5 with a 0.5-T Siemens Magnetom,
1 with a 1.0-T Siemens Magnetom, and 17 with a 1.5-T
Philips Gyroscan). Cranial MRI scans were performed in all
the NonDyG patients (30 with a 1.5-T Philips Gyroscan, and
10 with a 0.5-T Siemens Magnetom). T1-weighted images
were routinely obtained in the coronal, axial, and sagittal
planes using a repetition time (TR) of 450 to 550 msec and
an echo time (TE) of 25 to 40 msec. Proton density and
T2-weighted images were obtained in the coronal and axial
planes, using a T R of 1,500 to 2,100 msec and a TE of 30
and 60 msec, respectively. The localization and anatomical
extent of the structural lesions were judged according to both
477
neuroimaging findings and the appearance of the brain at the
time of the operation. Lesions were classified as focal if they
occupied less than one-third of a lobe. Larger lesions within
a lobe were considered lobar. When portions of more than
one lobe were involved, the lesion was classified as multilo-
bar. The central region (pre- and postrolandic cortex) was
considered as a separate lobe.
All patients included in this study underwent acute preex-
cision ECoG recordings. An ECoG recording apparatus con-
sisting of 16 carbon-ball electrodes from the same maker
(MNI Neuroelectronics Laboratory, Montreal, Canada) was
used in all three Centers. Interelectrode distance varied be-
tween 1.5 and 2.5 cm in a 4 X 4 matrix. Except for 4 patients
of the CDyG all also had postexcision ECoG recordings.
At least 10 minutes of recording were reviewed from each
examination. Preexcision ECoGs were evaluated for the pres-
ence and distribution of interictal spikes and for the presence
or absence of continuous or recurrent, rhythmic or quasi-
rhythmic trains of epileptic activity, of variable frequency and
configuration, unequivocally suggesting I/CEDs (see below).
The anatomical extent of the cortical areas displaying I/CEDs
was noted and later correlated with the extent of the struc-
tural lesion and of the areas displaying interictal ECoG spik-
ing. When I/CEDs or interictal spikes were recorded over
three or fewer adjacent ECoG electrodes within a lobe, they
were considered focal. If more than three adjacent electrodes
or any number of nonadjacent ECoG electrodes within a
lobe recorded I/CED or interictal spikes, they were consid-
ered lobar. If electrodes in more than one lobe displayed
I/CED or interictal spikes, these were classified as multilo-
bar. Completeness of excision of the epileptogenic cortex was
analyzed according to the ECoG, taking into account both
interictal spiking and I/CEDs. Comparing pre- and postexci-
sion records, resections were considered complete or partial.
Twenty of the 33 (60%) CDyG and 17 (42.5%) of the 40
NonDyG patients were operated o n under general anesthe-
sia. The remainder were operated on under local anesthesia
plus neuroleptoanalgesia (an association of a neuroleptic and
an opioid analgesic). The outcome of surgical treatment was
analyzed only for the CDyG, and classified as in our previous
report [20}, as follows: class A, seizure free, auras only, or
recurrence of seizures only on withdrawal of medication;
class B, reduction by greater than 90% of major seizures and
at least 75% of minor seizures with clear improvement in
social functioning; class C, reduction by greater than 50% of
major and minor seizures; class D , reduction by less than
50% of major seizures, irrespective of minor seizures; and
class E, no change in seizure frequency. Four patients in the
CDyG had a less than I-year follow-up, and in 3, outcome
information was not available. Mean follow-up for the other
27 patients was 4.3years, ranging from 1 to 15 years.
Statistical analyses consisted of contingency tables and x2
with continuity correction to correlate categorical variables.
Student’s t test was applied to compare means of two groups.
Results
Demographic und Anesthetic Data
There was no statistically significant difference in sex
distribution, duration of epilepsy, and type of anesthe-
sia between the two groups. Patients in the CDyG
478 Annals of Neurology Vol 37 No 4 April 1995
were significantly younger than patients in the Non-
DyG, both at epilepsy onset and at the time of the
operation (both p < 0.05).
Characterization of IICEDs
IICEDs were present in the preexcision ECoGs of 23
of the 34 patients (67%) in the CDyG (see below).
This highly epileptogenic activity was considered to be
present when one of the following patterns of rhythmic
epileptogenic activity was repeatedly observed in the
ECoG recordings: ( 1) repetitive electrographic seizures
(recruitinglderecruiting prolonged trains of rhythmic
activity), (2) repetitive bursting discharges (sudden
bursts or trains of spikes at 10 Hz or faster), of variable
duration, or (3) continuous or quasicontinuous rhyth-
mic spiking (rhythmic spikes or sharp waves at 1-8
Hz, for more than 10 seconds). In some patients, more
than one pattern was present.
REPETITIVE ELECTROGRAPHIC SEIZURES (RECRUITING~DERE-
CRUITING PATTERN). This morphologic pattern was
seen in 11 of the 23 patients (48%). Isolated spikes
progressively increased in frequency and rhythmicity
for several seconds, attained a frequency plateau
around 12 to 16 H t , and later slowed (Fig 1). This
sequence was followed by focal slowing or attenuation
of the recording. In 2 of the patients, while some re-
gions displayed this pattern, others showed quasicon-
tinuous, slower rhythmic spikes (see below).
REPETITIVE BURSTING PATTERNS. This pattern was ob-
served in 7 of the 23 patients (30%). High-frequency,
rhythmic polyspikes appeared suddenly, lasted for 5 to
10 seconds, and abruptly disappeared. Frequency var-
ied from 10 to 20 Hz, or faster (Fig 2A). There was
no change in background rhythms after the bursts. In
1 patient, the same electrodes alternated between this
pattern and repetitive electrographic seizures.
CONTINUOUS OR QUASICONTINUOUS RHYTHMIC SPIKING.
Prolonged trains of rhythmic 2- to 8-Hz spikes or sharp
waves were seen on the preexcision ECoG records of
8 patients (35%) (Fig 3). In 2, this pattern occurred in
some regions, while in other regions repetitive electro-
graphic seizures were seen. In 3 patients, spiking was
literally continuous, with an almost fixed periodicity
(Fig 4).
Relation of IICED on ECoG to Frequent Spikes or
Electrographic Seizures on EEG
Trains of rhythmic or quasicontinuous spikes or sharp
waves on scalp/sphenoidal EEG, or recurrent electro-
graphic seizures, were recorded in 15 of the 34 patients
(449%)in the CDyG. In 12 of these (80%), IICEDs
were also present on ECoG (see Fig 2A and B). There-
fore, in more than half the patients with IICEDs on
Fig I . Acute electrocorticographic recording in the referential
montage of a 12-year-old girl with a magnetic resonance im-
aging-negative, histologically proven dysplastic lesion in the left
occipital region. Note a recruitinglderecruitingepiteptogenicpat-
tern, characterizing an electrographic seizure, maximum ouer the
inferior occipital region. This pattern of IICED recurred many
times during the recording.
ECoG, a rhythmic spiking pattern was also present on
scalp/sphenoidal EEG. There was no correlation be-
tween the pattern of I/CEDs on ECoG with the epilep-
togenic pattern on EEG, and bursting discharges were
not observed in the latter. In no instance was a similar
pattern observed in the NonDyG.
Spatial Distribution of IICEDs
RELATION TO UNDERLYING PATHOLOGY. All morpholog-
ical patterns of I/CEDs were recorded from electrodes
overlying both macroscopic and microscopic dysplastic
cortex. Macroscopic abnormalities were identified by
correlation with imaging, by direct cortical visualiza-
tion, or both. In 5 patients, cortical areas displaying
IlCEDs on ECoG were labeled during the recording
and later histologically analyzed. In all 5, microscopic
dysplastic tissue was clearly more widespread than
could be predicted from imaging andlor visual inspec-
tion (Palmini A, unpublished observations, 1994).
I/ CEDs were recorded from electrodes overlying these
imaging-negative but microscopically dysplastic areas.
Since not all areas displaying I/CEDs were as carefully
analyzed in all patients, it cannot be ruled out that
this pattern may also have been recorded from some
Palmini et al: Epileptogenicity of Dysplastic Cortex
microscopically normal nondysplastic corticd regions,
but it clearly correlated with dysplastic cortex.
RELATION TO SPATIAL DISTRIBUTION OF THE MRI-VISIBLE
LESION. In 19 of the 23 patients (82%) of the CDyG
who displayed I/CEDs on ECoG, the spatial distribu-
tion of this activity was concordant with that of the
visible structural lesion. In 15 of these, both were focal
or lobar, and in 4 multilobar. A discordance in spatial
distribution was observed in only 4 patients, in whom
I/CEDs had a focal or lobar distribution, but the lesion
was more extensive or multilobar (Table 1).This spatial
colocalization was statistically significant at p < 0.05.
RELATION TO SPATIAL DISTRIBUTION OF THE INTERICTAL
ECOG SPIKING. Twenty-eight of the 34 patients (82%)
in the CDyG had multilobar interictal ECoG spiking,
while this was true for only 8 of the 40 patients (20%)
of the NonDyG ( p < 0.01).
Among the 23 patients of the CDyG whose ECoGs
displayed I/CEDs, when the latter were multilobar (4
patients) this coincided with the distribution of interic-
tal ECoG spiking. Of the remaining 19 with a focal
or lobar IlCED distribution, 16 (84%) had multilobar
interictal ECoG spiking. Although this comparison did
not reach statistical significance, it clearly showed that
interictal ECoG spiking was more widespread than
I/CEDs (see Table 1).
Finally, there was no statistically significant correla-
tion between the extent of the visible lesion and that
of the interictal spiking. Analyzing the same 23 patients
479
F i x 2. Seven-year-old girl with a right fronto-central dysplastic
lesion on magnetic resonance imaging. (A) Note repetitive bursts
of polyspikes, with variable amplitude and duration, recorded
diffusely from the right fronto-central regions on electrocorticog-
raphy. ( B ) Scalp electroencephalographic recording displaying
quasicontinuous sharp waves over the same regions.

480 Annals of Neurology Vol 37 No 4 April 1995


I-R
2-R
3 - R j
4-R-
5 -
6-R
7-R
10 - R
I1 - R
Fig 3. Acute electrocorticographic recording of a 4-year-old girl
with a dysplastic cortical abnormality involving the right
centro-temporo-parietal region on magnetic resonance imaging.
Note continuous, rhythmic, or semirhythmic spikes, recorded
from centro-temporo-parietal electrodes.
who had I/CEDs on ECoG, the spatial distribution of
the lesion was concordant with that of the interictal
ECoG spiking in 11, and discordant in 12. When it
was discordant, the spatial distribution of the interictal
spiking was always greater than the visible lesion (see
Table 1).
In contrast, the distribution of interictal ECoG spik-
ing tended to colocalize with the structural nondysplas-
tic lesion (NonDyG); i.e., in 32 of the 36 patients
(89%) with a focal or lobar lesion, interictal ECoG
spiking had a similar distribution. The latter was
multilobar in only 4 of the 36 patients (11%) with a
focal or lobar nondysplastic lesion, and in the 4 with a
multilobar lesion.
Comparison with Other Cortical Lesions
One o r more of the eiectrographic patterns of IICEDs
were present in the preexcision ECoG recordings of
23 of the 34 (67%) CDyG patients. This contrasted
sharply with the occurrence of similar activity in only
1 of 40 (2.5%) patients with other cortical lesions
(NonDyG) ( p < 0.001). The latter group was heavily
Palmini et al: Epileptogenicity of Dysplastic Cortex
T C PIS ECOG 3220
2
R 5
skewed toward slowly growing tumor patients, none
of whom displayed IICEDs. The only patient in the
NonDyG showing a pattern similar to I/CEDs was a
26-year-old woman with a history of slowly progressive
unilateral epileptic encephalopathy during childhood
and adolescence, accompanied by hemiparesis. Pathol-
ogy was compatible with a burned-out form of Rasmus-
sen’s encephalitis. The ECoG showed almost constant
spiking over the central region.
Relevance for Surgical Planning in Patients with
Cortical Dysplasia
To determine whether the identification of cortical re-
gions displaying I/CEDs on ECoG had an impact on
the surgical strategy in patients with CDyLs and intrac-
table epilepsy, we analyzed the 18 patients of the
CDyG who fulfilled the following criteria: (1) They
had I/CEDs on preexcision ECoG, (2) postexcision
ECoG was available for analysis, to assess whether cor-
tical regions displaying I/CED persisted or had been
removed, and (3) they had a greater than l-year post-
operative follow-up. Five patients with IiCEDs were
excluded from this analysis; 4 had a less than l-year
follow-up, and in the other, outcome information was
not available.
In 12 of the 18 patients, complete resection of the
cortical regions displaying I/CEDs was possible, as
481
Fig 4. Acute ele~trocorticogramof an 8-year-old boy with a
right central dysplastic lesion on magnetic resonance imaging.
Note continuous. rhythmic, or semirhythmic spiking in the pre-
central cortex. Two electrocardiographic (EKG) leads are shown,
to illustrate the striking similarity between the continuous epilep-
tic abnormality and EKG rhythmicity. I t is likely that an in-
trinszc ')acemaker" is also responsible for the continuous and
rhythmic abnormality of the dysplastic cortex.

Table I . ~sionl~lec~rographic
Correlations
Extent
IlCEDs Interictal Spiking

Extent Focal/Lobar Multilobar Focal/Lobar Multilobar Statistics

~

Lesion
Focalllobar 15 0 p < 0.05
Multilobar 4 4
Lesion
Focalllobar 3 12 NS
Multilobar 0 8
IlCEDs
Focalllobar 3 16 NS
Mu1tilobar 0 4
I/CEDs = ictal or continuous epileptogenic discharges; NS = nonstatistically significant (x2,p > 0.05).

482 Annals of Neurology Vol 37 No 4 April 1995

Table 2. Thirty-four Patients with Localized Cortical Dysplastic Lesions
Patient Age at Seizure Age at IICEDs on IICEDs on IICEDs on Lesion Lesion Outcome
No. Onset (yr) Surgery (yr) EEG Pre-Exc ECoG Post-Exc ECoG Location Excision” Class
1 1.5 10 Absent Present Disappeared Orbito-F MAJ C
2 10 19 Absent Absent N/A Lateral-F COMP D
3 2 19 Present Present Disappeared Lateral-F MAJ B
4 2 25 Absent Present Persisted F-Ct MAJ C
5 8 27 Absent Absent NIA Lateral-F MAJ Unknown
6 0.5 3 Absent Present Persisted Lateral-F MAJ C
7 6 13 Absent Present Disappeared Lateral-F MAJ B
8 0.5 3 Present Present Disappeared F-Ct-T COMP A
9 0.1 0.8 Present Present Disappeared Hemisph COMP Unknown
10 1 8 Present Present Persisted Diffuse-F MAJ C
11 1 21 Present Present Disappeared Ct MIN B
12 2 30 Present Absent N/A F-T MAJ Unknown
13 3 17 Absent Present Disappeared T-P MAJ B
14 0.5 4 Absent Present Unknown T-P-0 MAJ D
15 3 12 Present Present Persisted T-P-0 MIN C
16 16 23 Present Present Disappeared Ct MAJ B
17 19 20 Absent Absent N/A T-P-0 MIN <1 yr
18 2 38 Absent Absent N/A T-P MIN <1 yr
19 11 23 Absent Present Disappeared Ct MAJ A
20 15 22 Absent Present Unknown T-P-Ins MIN D
21 1 19 Present Present Disappeared Lateral-F COMP A
22 9 19 Present Absent NIA F-Ct MAJ B
23 5 24 Present Present Disappeared P MAJ B
24 14 25 Absent Absent NIA Post-T MIN C
25 10 18 Absent Absent NIA Post-T/P MIN D
26 8 30 Present Present Persisted Ct MIN D
27 5 13 Absent Present Persisted Post-T MIN D
28 4 11 Present Absent N/A F-Ct MAJ B
29 1 20 Absent Absent NIA Hemisph MIN D
30 0.5 5 Absent Absent N/A F MAJ D
31 1 7 Present Present Persisted F-Ct MIN D
32 1 13 Absent Present Disappeared F MAJ D
33 0.1 0.4 Present Present Disappeared F-Ct-P-0 MAJ <1 yr
34 2 8 Present Present Unknown Ct MAJ <1 yr
”See text for details.
Ct = central region; COMP = complete excision; ECoG = electrocorticography; EEG = electroencephalogram; Exc = excision; F = frontal
lobe; Hemisph = hemispheric (diffuse); I/CEDs = ictal or continuous epileptogenic discharges; Ins = insular region; MAJ = major excision
(>80%); MIN = minor excision (<SO%); N/A = not applicable; No. = number; 0 = occipital lobe; P = parietal lobe; T = temporal
lobe

shown by the disappearance of this pattern in postexci-
sion ECoG recordings. Nine of these 12 (75%) had a
good or excellent surgical outcome (class A or B) in
postoperative seizure control (Table 2). Conversely,
none of the 6 patients in whom I/CEDs persisted on
postexcision ECoG had a favorable outcome ( p <
0.01) (Fig 5).
To study the correlation between the extent of exci-
sion of cortical tissue displaying interictal spikes on
ECoG and surgical outcome, we analyzed the 23 pa-
tients with CDyLs who had both pre- and postexcision
ECoG recordings and a greater than 1-year postopera-
tive follow-up. Three of the 5 patients (60%) in whom
interictal spikes had completely disappeared from the
postexcision record had a favorable outcome, while the
same results were attained by 5 of the 18 patients
(28%) in whom spikes persisted in postexcision
ECoGs. This difference was not statistically significant
( p = 0.41).
Discussion
It would be unjustified to consider all three electro-
graphic patterns of IlCEDs described above as un-
equivocally ictal. The use of terms like “bursting” or
“continuous” epileptogenic discharges is prudent, at
least until more is learned about the microphysiologic
aspects of this hypersynchronous activity.
The recruiting/derecruiting pattern is clearly ictal
(see Fig 1). This corresponds to the classical mode of
initiation and evolution of seizure activity in both extra-
cranial and intracranial electrographic recordings of
partial seizures (28). The other two patterns, repetitive
bursting (see Fig 2A) and continuous or quasicontinu-
ous rhythmic spikes or sharp waves at slower frequen-

Palmini et al: Epileptogenicity of Dysplastic Cortex 483

14-15 **

A
V F Post ECoG 2980

2-3 \
3-4
5-6 1
6-7 d e &

15-16
B
F i g 5 . Pre- and po-stexcisionacute electrocorticographic (ECoGI
recording of a 25-year-old man with a right fronto-central dys-
plastic lesion on magnetic resonance imaging. Preexcision ECoG
(A, shows multzfical, apparently independent regions diplaying
I
repetitive electrographic seizures. Note persisteme of similar activ-
itji around central region in the postexcision recording (B). This
persistence of ictal or continuous epileptogenic discharges was as-
sociated with poor J urgical outcome.

484 Annals of Neurology Vol 37 No 4 April 1095

cies (see Fig 3) are less clearly ictal in nature but may
represent intermediate stages between interictal and
ictal states. They display several characteristics of the
ictal state, like hypersynchrony and rhythmicity, but
lack “postictal” slowing or attenuation usually seen fol-
lowing true electrographic seizures. In this regard they
are “ictal-like,” a term frequently used to describe the
same abnormalities in experimental studies in vitro
t291.
Some apparently contradictory findings on positron
emission tomography (PET) and single-photon emis-
sion computed tomography (SPECT) studies also sug-
gest a transitional ictal/interictal state. Even though nei-
ther this nor any other study has so far correlated the
ECoG pattern of spiking with PET or SPECT results,
a transitional ictal/interictal state is suggested by the
variable findings observed on these imaging studies in
relation to continuous spiking on EEG. Chugani and
colleagues [30] reported 4 children with no clinical
seizures but extremely frequent epileptogenic EEG
spikes recorded concomitantly with PET imaging ac-
quisition. Instead of the more typical interictal hypome-
tabolism, PET displayed interictal hypemetabolism. The
latter is usually considered a marker of ictal activity on
PET. O n the other hand, our Patient 15 had interictal
SPECT displaying hypoperjiusion over a dysp!.astic area
showing continuous spiking, as demonstrated by con-
comitant EEG monitoring. It is possible that the EEG/
ECoG patterns of I/CEDs reported here reflect de-
grees of epileptogenicity that fluctuate between the in-
terictal and ictal state, as also speculated by Chugani
and colleagues 130) to explain their PET findings. Not-
withstanding these remarks, it should be mentioned
that most PET studies performed in patients with corti-
cal dysplasia show interictal hypometabolism [3 1-33].
However, even though many of these patients were
operated on for intractable seizures, there are no data
regarding the occurrence of I/CED patterns on EEG
or ECoG. The 5 patients with focal PET hypometab-
olism and cortical dysgenesis, reported in detail by
Chugani and colleagues {31}, had frequent focal EEG
spiking but apparently not the “extremely frequent
spiking” found in patients with interictal PET hyper-
metabolism 130). In a study presenting data on ECoG
and PET in 7 children with cortical dysgenesis, Olson
and co-workers [ 3 2 ) concentrated on the spatial local-
ization of ECoG and PET abnormalities. The specific
issue of spiking frequency in relation to PET was not
addressed, although their Figure 3 shows a typical ex-
ample of what we called continuous or quasicontinuous
spiking at low frequency. Perhaps the low spiking fre-
quency had to do with the reported interictal PET focal
hypometabolism in that patient. The current evidence,
therefore, suggests that the specific pattern displayed
by functional imaging studies has to do with the state
of epileptogenic activation of the dysplastic tissue; i.e.,
Palmini et al: Epileptogenicity of Dysplastic Cortex
either focal hyper- or hypometabolism may be found
on interictal PET scans, although the latter occurs more
frequently.
It is believed that the interictal state is maintained
through strong inhibitory influences within the epilep-
tic focus [28). In most experimental models, when in-
hibition is decreased, the intracellular and field record-
ings show an increase in epileptic activity [34-361. All
three morphological patterns of I/CEDs suggest some
impairment of local inhibitory circuits, leading to the
prolonged trains of epileptic activity. The observation
by Ferrer and colleagues El51 that GABAergic in-
terneurons are indeed decreased in dysplastic cortical
regions, compared with adjacent nondysplastic cortex,
is therefore relevant. In their landmark report, they
have shown for the first time that there is a neurochem-
ical basis for an intrinsic and high degree of epilepto-
genicity in cortical dysplastic lesions. In a girl operated
on for intractable malignant partial motor seizures, the
authors compared the immunoreactivity of dysplastic
rolandic cortex with surrounding architecturally nor-
mal cortex using antibodies against parvalbumin and
calbindin 28K. The latter are intracellular calcium
binding proteins that reliably identify GABAergic in-
terneurons in cortical tissue. They demonstrated that
in dysplastic cortex there was a considerable reduction
in the number of GABAergic interneurons (local cir-
cuit neurons). In addition, Hoffman and associates [37]
in a recent elegant series of experiments showed that
reduction of GABAergic activity uncovered latent epi-
leptogenicity in a chronic model of neocortical epilep-
togenesis in vitro; previously “silent” slices started to
generate spontaneous discharges upon addition of bi-
cuculline to the perfusion medium. These findings, and
the present observation of I/CEDs recorded directly
from dysplastic cortex, suggest strongly that these le-
sions are intrinsically epileptogenic. A striking example
of this is shown by our Patient 34. During intraopera-
tive ECoG recording, continuous “needle-like” fast
spikes were recorded at a frequency of 1 to 3 Hz, in
an almost rhythmic fashion. Comparison with concomi-
tant electrocardiographic (EKG) recording (see Fig 4)
showed that the epileptic activity was independent
from, but closely resembled, the cardiac electric activ-
ity, which follows a known pacemaker. It is probable
that a somewhat similar mechanism of recurrent excita-
tion in the presence of decreased inhibition “paces”
the continuous or intermittent but usually rhythmic
activity that we termed UCED and that points to the
intrinsic epileptogenicity of the dysplastic cortex. Pre-
liminary observations also suggest that unlike epilepto-
genic neocortical tissue associated with other etiolo-
gies, dysplastic neocortical slices display spontaneous
bursting activity (Avoli M, personal communication,
1993).
Recurrent excitation is indeed a property of the nor-
485
mal functioning of the neocortex in mammals 1381.
If it is associated with the presence of spontaneously
bursting cells and decreased inhibitory activity, as we
have discussed above for dysplastic lesions, then the
three classical components of intrinsic epileptogenicity
are present [391. This property of intrinsic epilepto-
genicity is shared only by nonneocortical epileptogenic
tissue, notably hippocampal sclerosis 128).
In contrast, ECoG electrodes overlying other struc-
tural lesions associated with epilepsy show either no
epileptogenic activity or only sporadic spikes. Electrical
activity is usually decreased over such lesions 118). The
adjacent neocortex produces the spikes and shows in-
hibitory abnormalities [ 19). Neocortical tissue adjacent
to nondysplastic structural lesions such as tumors, arte-
riovenous malformations, or cysts is probably less epi-
leptogenic than cortical dysplastic tissue. This was sug-
gested by the significantly higher occurrence of
I/CEDs in the ECoG of patients with CDyLs compared
with those with other lesions. This difference is not
explained by anesthetic variables, or by the duration of
the epilepsy, which did not significantly differ between
both groups. Finally, the length of ECoG recordings
was also not different between both groups. Despite
the possible caveat related to the retrospective nature
of this study, we do not believe the use of our routine
ECoG recording protocol (10-minute recordings, the
same in all three institutions) had any impact on results.
I/CED patterns were either continuous or recurred so
often from the very start of the recordings that false-
negative findings in either group are highly unlikely.
Nevertheless, it is possible that more patients of the
NonDyG might have shown IlCEDs surrounding the
lesion with chronic subdural recordings. Though the
methodology of the present study could not provide
evidence for such abnormalities, we believe that our
findings in an acute surgical setting support the view
that dysplastic tissue is more epileptogenic than other
types of structural lesions or gliotic epileptogenic neo-
cortical tissue.
Taking into account the finding of I/CEDs on ECoG
and the presence of prolonged trains of rhythmic
spikes on scalp EEG (see Table 2 and {40}),26 of 34
patients (76%) had some ictal or continuous epilepto-
genic pattern. Therefore, it appears that the intrinsic
histological and neurochemical abnormalities of the
dysplastic cortex have a higher epileptogenic potential
than the “indirect epileptogenicity” associated with
other structural lesions. This possibility is also sup-
ported by recent clinical data. We have proposed 125)
that patients with cortical dysplastic lesions have a
greater tendency to intractable seizures and a more
frequent history of partial or generalized status epilep-
tics compared with patients with other lesions. Status
epileptics occurred in 30% of patients with cortical
dysplasia compared with 3 to 49% in series of patients
486 Annals of Neurology Vol 37 No 4 April 1005
with supratentorial neoplasms and epilepsy E41). Re-
porting on ictal SPECT in patients with CDyLs, Kuz-
niecky and colleagues 142) showed that ictal cerebral
blood flow was focally increased in the regions where
MRI showed cortical dysplastic abnormalities, thus
adding support to the contention that dysplastic tissue
is intrinsically epileptogenic.
A final point concerns the relevance of identifying
I/CEDs intraoperatively for determining the ideal sur-
gical strategy for treatment of intractable seizures in
patients with CDyLs. We have previously shown that
the only reliable predictor of favorable surgical out-
come was the extent of removal of the visible lesion,
detected either by MRI or by direct cortical inspection
[20). Extent of removal of the irritative zones deline-
ated by both EEG and ECoG could not predict out-
come 120). Since it has been demonstrated that micro-
scopic dysplastic abnormalities can occur in the context
of a normal MRI and of macroscopically normal cortex
116, 431, one can never be sure that resecting the visi-
ble lesion means excising all dysplastic tissue. This may
even explain the number of patients who had unfavor-
able outcomes despite major removal of the visible
dysplastic area in that study 120). Functional imaging
studies have identified dysplastic tissue 116, 30-311,
and may prove useful in demonstrating dysplastic areas
surrounding the MRI-defined lesion. However, we
have shown that I/CEDs recorded during acute ECoG
may be important in determining the amount of tissue
to be excised. Three-fourths of the patients in whom
all cortical areas displaying I/CEDs were excised had a
favorable surgical outcome. When highly epileptogenic
tissue remained in situ, as indicated by the persistence
of I/CEDs on postexcision ECoG recordings, outcome
was uniformly unfavorable. We believe it can now be
reliably stated that the best surgical strategy for treat-
ment of medically refractory seizures associated with
CDyLs is (1) to remove as much as possible of the
visible lesion, and (2) to remove as much as possible
of the cortical areas displaying I/CEDs on acute ECoG.
Fortunately, we have demonstrated that both have a
strong tendency to colocalize. Resecting the tissue gen-
erating IJCEDs may make the difference between a
successful and an unfavorable outcome in the surgery
of epilepsy associated with cortical dysplasia.
References
1. Spencer DD, Spencer SS, Mattson RH, Williamson PD. Intrace-
rebral masses in patients with intractable partial epilepsy. Neu-
rology 1984;34:432-436
2. Cascino GD, Kelly P, Hirshorn KA, et al. Stereotactic resection
of intraaxial cerebral lesions in partial epilepsy. Mayo Clin Proc
199O;65:1053-1060
3. Fish D, Andermann F, Olivier A. Surgical strategies in patients
with complex partial seizures and small posterior temporal or
extratemporal structural lesions. Neurology 1991;41:1781-
1784
4. Awad I, Rosenfeld J, et al. Intractable epilepsy and structural
 lesions of the brain: mapping, resection strategies, and seizures
outcome. Epilepsia 1991;32: 179- 186
5. Boon P, Williamson P, Fried I, et al. Intracranial intraaxial,
space-occupying lesions in patients with intractable partial sei-
zures: an anatomoclinical, neuropsychological, and surgical cor-
relation. Epilepsia 1991;32:467-476
6. Barkovitch AJ, Chuang SH, Norman D. MR of neuronal migra-
tion anomalies. AJNR 1987;8:1009-1017
7. Kuzniecky R, Berkovic S, Andermann F, et al. Focal cortical
myoclonus and rolandic cortical dysplasia: clarification by mag-
netic resonance imaging. Ann Neurol 1988;23:317-325
8. Palmini A, Andermann F, Tampieri D, Robitaille Y. Neuronal
migration disorders: a contribution of modern neuroimaging
techniques to the etiologic diagnosis of epilepsy. Can J Neurol
Sci 1771;18:580-587
9. Palmini A, Andermann F, Olivier A, et al. Focal neuronal migra-
tion disorders and intractable partial epilepsy: a study of 30
patients. Ann Neurol 1991;30:741-749
10. Guerrini R, Dravet C, Raybaud C, et al. Epilepsy and focal
gyral anomalies detected by MRI: electroclinico-morphological
correlations and follow-up. Dev Med Child Neurol 1992;34:
706-7 18
11. Brodtkorb E, Nilson G , Smevik 0,Rmck P. Epilepsy and anom-
alies of neuronal migration: MRI and clinical aspects. Acta Neu-
rol Scand 1992;86:24-32
12. Robitaille Y, Rasmussen T, Dubeau F, et al. Histopathology of
nonneoplastic lesions in frontal lobe epilepsy. In: Chauvel P,
Delgado-Escueta A, Halgren E, Bancaud J, eds. Frontal lobe
seizures and epilepsies. New York: Raven Press, 1992:499-
513
13. Taylor DC, Falconer MA, Bmton CJ, CorsellisJAN. Focal dys-
plasia of the cerebral cortex in epilepsy. J Neurol Neurosurg
Psychiatry 197 1;34:369-387
14. De Rosa M, Secor D, Barsom M, et al. Neuropathologic find-
ings in surgically treated hemimegalencephaly: immunohisto-
chemical, morphometric, and ultrastructural study. Acta Neuro-
pathol (Bed) 1992;84:250-260
15. Ferrer I, Pineda M, Tallada M. Abnormal local-circuit neurons
in epilepsia partialis continua associated with focal cortical dys-
plasia. Acta Neuropathol (Bed) 1992;83:647-652
16. Desbiens R, Berkovic S, Dubeau F, et al. Life-threatening focal
status epilepticus due to occult cortical dysplasia. Arch Neurol
1993;50:695-700
17. Palmini A, da Costa JC, Andermann F, et al. Focal electro-
graphic ictal activity during acute cortical recording over dysplas-
tic lesions in humans. Epilepsia 1992;33:75 (Abstract)
18. Pilcher W, Silbergeld D, Berger M, Ojemann G. Intraoperative
electrocorticography during tumor resection: impact on seizure
outcome in patients with gangliogliomas.J Neurosurg 1993;78:
89 1-702
19. Haglaud M, Berger M, Kunkel D, et al. Changes in gam-
ma-aminobutyric acid and somatostatin in epileptic cortex asso-
ciated with low-grade gliomas. J Neurosurg 1992;77:209-
2 16
20. Palmini A, Andermann F, Olivier A, et al. Focal neuronal migra-
tion disorders and intractable partial epilepsy: results of surgical
treatment. Ann Neurol 1991;30:750-757
21. Kitasato H, Murayama K. An analysis of the change in K-
permeability on depolarization in terms of affinity and numbers
of total channels. J Theor Biol 1975;51:181-200
22. Sarnat H. Cerebral dysgenesis: embryology and clinical expres-
sion. New York: Oxford University Press, 1992
Palmini e t al: Epileptogenicity of Dysplastic Cortex
23. Perot P, Weir B, Rasmussen T. Tuberous sclerosis: surgical ther-
apy for seizures. Arch Neurol 1966;15:498-506
24. Andermann F, Olivier A, Melanson D, Robitaille Y. Epilepsy
due to focal cortical dysplasia with macrogyria and the forme
fruste of tuberous sclerosis: a study of 15 patients. In: Wolf P,
Dam M, Janz D, Dreifuss FE, eds. Advances in epileptology,
vol 16. New York: Raven Press, 1987:35-38
25. Palmini A, Gambardella A, Andermann F, et al. Operative strat-
egies for patients with cortical dysplastic lesions and intractable
epilepsy. Epilepsia 1994;35(suppl 6):S57-S7 1
26. Milner B. Psychological aspects of focal epilepsy and its neuro-
surgical management. In: Purpura D, Penry K, Walter R, eds.
Neurosurgical management of the epilepsies. New York: Raven
Press, 1975:299-32 1
27. Jones-Gotman M. Evaluation: testing hippocampal function. In:
Engel J Jr. ed. Surgical treatment of the epilepsies. New York:
Raven Press, 1987:203-211 (Commentary)
28. Engel J Jr. Seizures and epilepsy. Philadelphia: FA Davis, 1989
29. Hablitz JJ. Spontaneous ictal-like discharge and sustained poten-
tial shifts in the developing rat neocortex. J Neurophysiol 1987;
58:1052-1065
30. Chugani HT, Shewmon DA, Khanna S, Phelps ME. Interictal
and postictal focal hypermetabolism on positron emission to-
mography. Pediatr Neurol 1993;9:10-15
31. Chugani HT, Shields WD, Shewmon DA, et al. Infantile
spasms: I. PET identifies focal cortical dysgenesis in cryptogenic
cases for surgical treatment. Ann Neurol 1990;27:406-4 13
32. Olson DM, Chugani HT, Shewmon DA, et al. Electrocortico-
graphic confirmation of focal positron emission tomographic ab-
normalities in children with intractable epilepsy. Epilepsia 1990;
311731-739
33. Lee N , Radtke R, Gray L, et al. Neuronal migration disorders:
positron emission tomography correlations. Ann Neurol 1994;
35:290-297
34. Jensen M, Yaari Y. The relationship between interictal and ictal
paroxysms in an in vitro model of focal hippocampal epilepsy.
Ann Neurol 1988;24:591-598
35. Gale K. Progression and generalization of seizure discharge:
anatomical and neurochemical substrates. Epilepsia 1788;
29(suppl):S15-s34
36. Da Costa JC, Russo R, Guillermo G, Velluti J. Seizure-like
events in the in vitro turtle brain. Epilepsia 1992;33:37 (Ab-
s trac t)
37. Hoffman SN, Salin PA, Prince DA. Chronic neocortical epilep-
togenesis in vitro. J Neurophysiol 1994;71:1762-1773
38. Douglas RJ, Martin KAC. Neocortex. In: Shepherd GM, ed.
The synaptic organization of the brain. 3rd ed. New York: Ox-
ford University Press, 1990:389-438
39. Prince DA. Physiological mechanisms of focal epileptogenesis.
Epilepsia 1985;26(suppl 1kS3-S 14
40. Gambardella A, Palmini A, Dubeau F, et al. Electroencephalo-
graphic patterns in epileptic patients with neuronal migration
disorders. Epilepsia 1993;34:127 (Abstract)
4 I . Leppik IE. Status epilepticus. In: Porter RJ, Theodore WH, eds.
Epilepsy. Philadelphia: WB Saunders, 1986:633-644
42. Kuzniecky R, Mountz J, Wheatley G, Morawetz R. Ictal single-
photon emission computed tomography demonstrates localized
epileptogenesis in cortical dysplasia. Ann Neurol 1093;34:627-
63 1
43. Chugani H, Shewmon A, Shields W, et al. Surgery for intracta-
ble infantile spasms: neuroimaging perspectives. Epilepsia 1993;
34:764-771
487