Epileptic Disorders
MENUMRI morphological and volumetric study of the cingulate gyrus and its relevance in partial epileptic patients Volume 5, numéro 2, June 2003
Illustrations
Auteur(s) : Stéphane Kremer1, 2, Marc Braun2, Philippe Kahane1, F. Guillemin3, Jean-François Le Bas4, Alim Louis Benabid1
1. INSERM U318, Department of Neurosciences, University
Hospital, Grenoble, France.
2. Neuroradiology Department, University Hospital, Nancy,
France.
3. Clinical Epidemiology and Evaluation Department, Nancy,
France.
4. Neuroradiology Department, University Hospital, Grenoble,
France
Received December 9, 2002 ; Accepted April 25, 2003
For more than a decade, MRI has proved to be an essential
examination in presurgical management of patients with medically
intractable partial seizures [1]. In almost 90% of the cases, it
shows space-occupying lesions or morphological abnormalities, which
can help to delineate the epileptogenic area [2]. In temporal lobe
epilepsy, the most frequent form of drug resistant epilepsies (78%
in the surgical series) [3], hippocampal sclerosis is present in
70% of the cases [3, 4], and seems to be a reliable marker of the
side, and even of the site of the epileptic discharges
[1].
Since the cingulate gyrus is connected to several limbic
structures through the Papez pathways [5-7], and since it can be
involved in the propagation, or in the genesis of partial epileptic
seizures [8, 9], it could be interesting to look for subtle MRI
abnormalities in the cingulate gyrus, such as atrophy, which could
be indicative of its implication in the epileptogenic
area.
However, the cingulate gyrus is a complex anatomical structure,
limited by several sulci or sulcus portions, with marked
morphological variability, as shown by Ono et al. [10],
explaining the difficulties in its delineation (figure 1).
The aims of this work were:
First to describe the sulcal limits of the cingulate gyrus, trying
to define a "normalised cingulate gyrus".
Secondly, by using these limits, to look for cingulate gyrus
atrophy, using volumetric MRI analysis, in patients with
involvement of the cingulate gyrus in the epileptogenic area, as
shown by stereotactic intracerebral EEG (SEEG).
Patients and methods
Patients
MRI examinations were performed for 20 patients
(9 females and 11 males, age:
31.2 ± 9.39 years), suffering from partial
drug-resistant epilepsy, and for 20 healthy volunteers
(10 females and 10 males, age: 31.8 ±
7.7 years).
In the epileptic group, two were left-handed and 18 were
right-handed, and in the volunteer group, 10 were left-handed
and 10 were right-handed, as assessed by the Edinburgh
inventory. In the epileptic group, the Wada-test confirmed a left
hemispheric dominance for language in the 18 right-handed
patients, and a right hemispheric dominance in the two left-handed
patients.
Epileptic patients were included based upon the following
criteria:
– adults more than 18 years old;
– negative history of psychiatric disorders;
– no obvious cerebral morphological lesion in the cingulate gyrus
or in the neighboring structures on MRI;
– cingulate gyrus recording with at least one intracingulate
electrode, during the intracerebral electrode recording
procedure;
– planned and accepted surgical treatment of their epilepsy.
They were divided into three groups according to the involvement
of the cingulate gyrus in the genesis of seizures, as defined by
intracerebral recordings:
– CG 1 comprised five patients whose seizures clearly involved
the cingulate gyrus, and which was therefore considered as part of
the epileptogenic zone, thus justifying its resection (none of the
resected specimens showed any specific’ histological changes inside
the CG)
– CG 2 comprised seven patients in whom the cingulate gyrus
was only part of the discharges spread (no surgical resection).
– CG 3 comprised eight patients whose cingulate gyrus was not
involved at all.
Informed consent was obtained from the volunteers, who were
included in the study on the following criteria:
– adults more than 18 years old;
– negative history of neurologic and/or psychiatric disorders;
– no evident abnormality in the cingulate gyrus or
surrounding structures on the MRI.
Neuroimaging
All MR studies were performed at 1.5 T on a Gyroscan unit
(Philips Medical Systems): axial T1w 3D Gradient Echo acquisitions,
thickness: 1.5 mm, reformatting in all planes.
The sulci (anterior paraolfactory sulcus, cingulate sulcus, and
subparietal sulcus) were described on these MRI and the
intrasulcal gray matter was segmented in a semi-automatic manner
using the "Surgiscope Scopeplan" software by Elekta on a
Hewlett-Packard workstation.
The limits were drawn manually in the frontal plane, cuts were
apart and instant checking was available in the other planes
simultaneously visualized. These measurements were performed on
both cerebral hemispheres of the epileptic patients and normal
volunteers.
The Scopeplan software starting from the presented limits,
calculated the volume (cc) of the cingulate gyrus.
Morphological study
The anterior limit of the cingulate gyrus is formed by the
anterior paraolfactory sulcus and the lamina terminalis, the
superior limit by the cingulate sulcus, then the subparietal sulcus
and more posteriorly by the anterior part of the antecalcarine
sulcus, the inferior margin by the callosal sulcus, and the
posterior border by the isthmus of the cingulate gyrus (figure 1). We studied the
major sulci that showed, in agreement with Ono’s description [10],
important morphological variations, namely the cingulate sulcus,
the subparietal sulcus and the anterior paraolfactory sulcus.
For the cingulate sulcus the following features were noted:
– the position of its anterior end, according to Ono’s
classification [10] (figure 2);
– its connection with the superior rostral sulcus;
– the presence of interruptions, defined as a rupture of the
continuity of the sulcus, visible on all MRI planes (figure 3);
– double parallel patterns, defined as a division of the
cingulate sulcus in 2 parallel subsulci, not connected with
each other (figure
3);
– The presence of an intracingulate sulcus, defined as a
sulcus starting at the anterior part of the corpus callosum and
joining the middle part of the cingulate sulcus (figure 3);
For the subparietal sulcus (figure 4): according to
Ono’s classification [10], it was postulated that the subparietal
sulcus has, most of the time, an H-pattern, defined by a horizontal
sulcus prolonging posteriorly the cingulate sulcus, and two
upwardly-oriented side-branches, and two downwardly-oriented
side-branches. The following features were noted:
– the number of side branches;
– the connection with the posterior end of the cingulate sulcus
and the route of connection;
– double parallel patterns;
– atypical patterns.
For the anterior paraolfactory sulcus, the following features were
studied:
– whether or not it was distinguishable;
– the connection of the anterior paraolfactory sulcus with the
cingulate sulcus.
These analyses were performed blind on both cerebral hemispheres
of each patient, by two independent observers (SK, MB).
Statistical analysis was performed using the Chi square test or
Fisher exact test, a P value of 0.05 was regarded as
significant. Due to the small number of patients in subgroups, the
different subgroups were compared as follows: epileptic hemispheres
of CG 1 (patients who presented seizures which involved the
cingulate gyrus) with those of CG 2 + 3 (patients in whom
the cingulate gyrus was only part of the discharges spreading or
was not involved at all), CG 1 with volunteers, CG
1 + 2 + 3 with volunteers, for the
presence of the different anatomical variations.
Inter-rater reliability was evaluated by the kappa index.
Volumetric analysis
The cingulate gyrus was measured after segmentation of the cortical grey matter, following the different sulci as anatomical limits defined above.
These analyses (figure 5) were performed blind on both cerebral hemispheres of each patient, by two independent observers (SK, MB), each one unaware of the results of the other observer.
Statistical analysis was performed using the Mann and Whitney U-test. The average volume of the cingulate gyrus was compared in the different subgroups (epileptic hemispheres of CG 1 (patients who presented seizures which involved the cingulate gyrus), with CG 2 + 3 (patients in whom the cingulate gyrus was only part of the discharges spreading or was not involved at all), CG 1 with volunteers, CG 1 + 2 + 3 with volunteers).Results
Morphological study
1) Inter-rater reliability
The inter-rater reliability was excellent, with a kappa index superior to 0.9 in all cases.
2) Statistical analysis
There was no significant statistical difference between the different subgroups, with the exception of the double parallel pattern that showed significantly more interruptions of the external portion in the volunteer group, than in the group CG 1 + 2 + 3 (P = 0.019, Chi square test). Nevertheless, these results allowed us to pool the 80 hemispheres, for the description of the sulcal variability.
3) Description of the sulcal variations (table 1-3) (figure 6 and 7)
Table 1. Cingulate sulcus: morphological description.
Cingulate sulcus | N | % |
---|---|---|
Position anterior end: a | 68 | 85 |
Position anterior end: b | 9 | 11 |
Position anterior end: c | 3 | 4 |
Double parallel pattern | 33 | 41 |
Single sulcus pattern | 47 | 59 |
No interruption (single sulcus pattern) | 29 | 62 |
One interruption | 14 | 30 |
Two interruptions | 4 | 8 |
No interruption internal sulcus (double parallel pattern) | 31 | 94 |
One interruption | 2 | 6 |
No interruption external sulcus (double parallel pattern) | 19 | 58 |
One interruption | 14 | 42 |
Intracingulate sulcus | 16 | 20 |
Cingulate sulcus-superior rostral sulcus connection | 8 | 10 |
Table 2. Subparietal sulcus: morphological description.
Subparietal sulcus | N | % |
---|---|---|
Cingulate sulcus-subparietal sulcus connection | 48 | 60 |
Route of connection antero-inferior sidebranch (Ant <)-marginal ramus (MR) | 21 | 44 |
Ant > -MR | 23 | 48 |
Horizontal branch of the H-MR | 4 | 8 |
Double parallel pattern | 3 | 4 |
Single sulcus | 72 | 90 |
Atypical | 5 | 6 |
Upwardly oriented sidebranches 1 | 11 | 14 |
Two | 57 | 71 |
Three | 7 | 9 |
Atypical | 5 | 6 |
Downwardly oriented sidebranches 1 | 7 | 9 |
Two | 62 | 77 |
Three | 6 | 8 |
Atypical | 5 | 6 |
H pattern | 46 | 57 |
Table 3. Anterior paraolfactory sulcus: morphological description.
Anterior paraolfactory sulcus | N | % |
---|---|---|
Anterior paraolfactory sulcus distinguishable | 56 | 70 |
Anterior paraolfactory sulcus not distinguishable | 24 | 30 |
Connection anterior paraolfactory sulcus- cingulate sulcus | 34 | 61 |
Volumetric study
1) Inter-rater reliability
The inter-rater reliability was excellent, with a intraclass correlation coefficient rho = 0.92 in the volunteer group and rho = 0.93 in the epileptic group.
2) Statistical analysis
No statistically significant difference was found between the mean volume in CG 1 (12.58 ± 2.84) and CG 2 + 3 (12.56 ± 2.57) (P = 0.89), between the mean volume in CG 1 and the volunteer group (12.56 ± 2.56) (P = 0.75), or between the mean volume in CG 1 + 2 + 3 (12.46 ± 1.94) and the volunteer group (P = 0.83).
The volumetry of the cingulate gyrus does not distinguish the different subgroups from each other.
Discussion
Description of the sulcal variability
In the double-parallel pattern of the cingulate sulcus, there were significantly more interruptions of the external portion in the volunteer group, than in the group CG 1 + 2 + 3 (P = 0.019, Chi square test), but there is no pathophysiological explanation for this result. No anatomical reason may explain such a difference.
The description of the sulcal variability was similar to that of other authors [10-12]. The few differences are related to a different definition of the sulci or to methodological differences.
The main differences in definition concerned the double parallel patterns of the cingulate sulcus described by Ono [10]. Actually, Paus [12] considers the external sulcus as the paracingulate sulcus and describes it as present in 85% of right-handed cases and in 92% of left-handed cases. But in 48% of the right-handed cases and 38% of the left-handed cases it consisted of only a few vertical branches [12], a pattern that has been excluded from the current study, considering only the horizontal sulci parallel to the cingulate sulcus, referred to as the proeminent paracingulate sulcus by Paus [12]. The vertical branches belong to other frontal lobe structures and not to the cingulate gyrus per se.
The intracingulate sulcus, present in 20% of the cases in the current study, corresponds to Paus, intralimbic sulcus, and is found in 6% on the right side and in 4% on the left side [11, 12]. This difference could be explained by the recruitment difference between the two studies, as Paus’s study pooled 247 subjects. Moreover this sulcus is not described by Ono [10].
Other differences may be explained by the properties of the MRI. Actually, more connections were found between the cingulate sulcus and the subparietal sulcus than by Ono [10] (table 5, Ono – 36% on the right side and 28% on the left side). MRI displays very thin slices (1.5 mm), and simultaneous multiplanar views allowing study of the depth of the sulci, detecting connections under the brain surface.
However, MRI is not suitable for a study of very superficial sulci, as they disappear within the slice thickness. This is, for example, the case for the anterior paraolfactory sulcus, described by Ono in 92% of the cases on the right side and in 72% on the left side, but in 20% on the right side and in 16% on the left side in the current series, it is represented by a smooth depression, not visible on MRI [10].
Another difference between this study and previous studies [10-12], is that the sulcal variability in the right hemispheres has not been compared to the left hemispheres. Paus has described significant differences between the two hemispheres in the cingulate sulcus [11-12].
However, in most of the studies, only right handed subjects were included. The series of subjects in this study was constituted by right and left handed subjects, and the cerebral dominance for language was assessed by an Amobarbital test, which can be considerate as the gold standard, only in the epileptic group. Cerebral dominance was assessed only by Edimburgh test in the volunteer group.
Methodological aspect
The cingulate gyrus is a complex anatomical structure. It
presents great sulcal variability, and in some portions no sulcal
limit is found, for example at the isthmus of the cingulate gyrus,
or in the case of an interruption in a sulcus or between two sulci.
The definition of a "normalised cingulate gyrus", based only on
sulcal limits is not possible, and it is necessary to define
arbitary limits, based on previous studies [13-15].
There were difficulties also in the double parallel pattern of the
cingulate sulcus in defining the cingulate gyrus limits. Whether
the area 32 of Broadmann, situated beween the two branches of
the cingulate sulcus, belongs to the cingulate cortex remains a
questionnable issue. Pandya, Room and Albanese [13, 16, 17]
consider, on studies based on the area 32 connections, that it
does not belong to the anterior cingulate gyrus cortex, but to the
prefrontal cortex. Talairach and Vogt [15, 18, 19] consider, based
on of cytoarchitectonic studies, that it represents a transitional
fronto-limbic cortex, though belonging to the cingulate gyrus. This
study is in agreement with these last authors, and we considers the
external branch of the cingulate sulcus as the superior limit of
the cingulate gyrus.
Moreover, the marginal ramus and the secondary, upwardly vertical
branches of the cingulate sulcus and of the subparietal sulcus were
excluded from the delimitation of the cingulate gyrus. Actually,
the marginal ramus penetrates the parietal lobe and does not
contain cingulate cortex. The same reasons apply to the secondary
branches, penetrating the parietal lobe or the frontal lobe.
The last limitation to this study, was the number of patients
showing implication of the cingulate gyrus in the genesis of the
epileptic seizures. Only 5 patients justifiing the resection
of the cingulate gyrus, were included in the study, and it is
statistically difficult to show small losses of volume in the
cingulate gyrus.
Volumetry of the cingulate gyrus in the epileptic patient
There was no difference in cingulate gyrus volume between the
different subgroups of patients and the healthy volunteers. The
volumetry of the cingulate gyrus did not allow us to distinguish
the group in whom the cingulate gyrus is involved in the genesis of
the seizures, from the other groups. There was no atrophy of the
cingulate gyrus in the epileptic group, unlike the hippocampal
sclerosis, which is found in 72% of the epilepsies involving the
temporal lobe [3].
Different hypotheses have been proposed to explain the appareance
of hippocampal sclerosis, but the most frequently cited is a past
history of convulsions, which could have been due to fever or to a
temporal malformative lesion [3, 4, 20].
Moreover, it has been demonstrated that epileptic seizures induced
by kindling, produced hippocampal lesions, and that hippocampal
neurons are particularly vulnerable to repeated seizures, compared
to other neuronal populations, which, when submitted to the same
treatment did not demonstrate any degeneration [20, 21]. This leads
to the hypothesis that cingulate neurons are less vulnerable to
repeated seizures than the hippocampic neurons. The fact that
repeated seizures induce atrophy seems to be a specific property of
hippocampal neurons, and of structures directly connected to the
hippocampus, such as the fornix, the mamillary bodies and the
entorhinal cortex [1, 22-26]. o