• Epilepsy;
  • Temporal lobe;
  • Stereoelectroencephalography;
  • Seizure;
  • Semiology


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments

Summary: Purpose: The International League Against Epilepsy (ILAE) classification distinguishes medial and neocortical temporal lobe epilepsies. Among other criteria, this classification relies on the identification of two different electroclinical patterns, those of medial (limbic) and lateral (neocortical) temporal lobe seizures, depending on the structure initially involved in the seizure activity. Recent electrophysiologic studies have now identified seizures in which medial and neocortical structures are both involved at seizure onset. The purpose of the study was therefore to study the correlations of ictal semiology with the spatiotemporal pattern of discharge in temporal lobe seizures.

Methods: The 187 stereoelectroencephalography-recorded seizures from 55 patients were analyzed. Patients were classified into three groups according to electrophysiologic findings: medial (M; seizure onset limited to medial structures, n = 24), lateral (L; seizure onset limited to lateral structures, n = 13), and medial-lateral (ML; seizure onset involving both medial and lateral structures, n = 18). Clinical findings were compared between groups.

Results: Initial epigastric sensation, initial fear, delayed oroalimentary and elementary upper limb automatisms, delayed loss of contact, long seizure duration, and absent or rare secondary generalizations were associated with M seizures. Initial auditory illusion or hallucination, initial loss of contact, shorter duration of seizures, and more frequent generalizations were associated with L seizures. Initial epigastric sensation, initial loss of contact, early oroalimentary and verbal automatisms, and long duration of seizures were associated with ML seizures.

Conclusions: Although the syndrome of mesial temporal epilepsy is now relatively well defined, our findings support the idea that the organization of temporal lobe seizures may be complex and that different patterns exist. We demonstrate three distinct patterns, characterized by both semiologic and electrophysiologic features. This distinction may help to define better the epileptogenic zone and the subsequent surgical procedure.

Since the early work from the Montreal and Paris schools using a number of approaches, including depth electrodes and cortical stimulation studies (1,2), the introduction of video-EEG analysis has resulted in a better knowledge of the phenomenology of temporal lobe seizures (3) and its distinction from that of complex partial seizures of extratemporal origin (4,5). Some early works have attempted to correlate ictal symptoms with several anatomoelectroclinical seizure subtypes (1,6,7). Subsequently, the International Classification of Epileptic Syndromes (8) made a distinction between medial and lateral temporal lobe epilepsies. Among other criteria such as medical history and pathologic data, this classification relies on the identification of two electroclinical patterns, that of medial (limbic) and lateral (neocortical) temporal lobe seizures (TLS) depending on the structure initially involved in the seizure activity. This classification has prompted extensive studies of the two corresponding syndromes (9–17). Some clinical features have been shown to be characteristic when comparing patients with medial temporal lobe seizures (MTLS) and those with lateral temporal lobe seizures (LTLS). A history of febrile convulsions, an ictal epigastric sensation, and early oral automatisms were more frequently associated with MTLS, whereas sensory hallucinations were more often associated with LTLS (16).

However, from an anatomic point of view, dense interconnections exist between the limbic and neocortical regions (18). Moreover, TLS related to a widespread epileptogenic zone involving both the limbic and neocortical structures have long been described, as reflected in earlier classifications (1,7,19). Recently, by using depth-EEG recordings and signal processing, we showed that the epileptogenic zone in TLE better corresponded to a network organization (20,21) and that this neural network sometimes involved in both medial (limbic) and lateral (neocortical) structures, thus defining a third subtype of seizures. This third subtype has never been individually characterized in previous semiologic studies comparing medial and neocortical TLE.

The purpose of the present study was to determine how the anatomic localization of the initial discharge influences the ictal semiology, by studying the correlations between clinical characteristics and the three electrophysiologic subtypes of TLS: medial, lateral, and medial-lateral.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments

Patient selection

We retrospectively studied 55 patients with medically intractable unilateral TLE investigated with stereoelectroencephalography (SEEG). They were selected from a group of 132 consecutive patients undergoing evaluation for surgical treatment of intractable TLE from 1994 to 1997 (Rennes) and from 1997 to 2001 (Marseille). All had prior comprehensive evaluation including a detailed history and neurologic examination, neuropsychological testing, routine magnetic resonance imaging (MRI) study, interictal and ictal (when available) single-photon emission computed tomography (SPECT), and video-EEG recordings of seizures. At the end of this first noninvasive phase, patients were separated into two groups: (a) if all the localization criteria were consistent with a pure clearly lateralized MTLE, patients were operated on (anterior temporal lobectomy or amygdalohippocampectomy); or (b) if discrepancies between the existence of hippocampal atrophy or other lesions and electroclinical and/or SPECT data occurred, SEEG was carried out to define the contribution (if any) of other temporal or extratemporal areas to the epileptogenic zone. In case of hippocampal atrophy, SEEG was indicated if a bitemporal epilepsy was suspected, and/or if semiology suggested a more widespread involvement of temporal lobe and particularly to determine the superior (superior temporal gyrus, insula) and posterior limits (posterior hippocampus and parahippocampal gyrus, temporo-parieto-occipital junction) of the epileptogenic zone.

Stereoelectroencephalography (SEEG)

Multiple-lead (10–15) intracerebral electrodes were orthogonally implanted according to Talairach's stereotactic method (22,23). The implantation scheme was based on a standardized approach, which was adapted individually according to the hypothesis of localization of the epileptogenic zone determined from noninvasive data and the individual anatomic constraints. These electrodes allowed sampling of both medial (amygdala, anterior and posterior hippocampus, entorhinal cortex) and lateral (temporal pole, superior, middle and inferior temporal gyri, temporo-occipital junction, temporoparietal junction) temporal structures, as well as extratemporal structures in some cases (Fig. 1).


Figure 1. Typical example of depth electrode implantation for stereoelectroencephalography (SEEG) temporal lobe epilepsy. A: Lateral view of all depth electrodes superimposed on a 3D image (created using Anatomist) (19,20). B: Coronal view of a preoperative plan of depth electrodes routes superimposed on a T1 image. The selected slice shows the position of the electrode 1 (in green), whereas the other electrodes are shown by transparency. Only a few electrodes are displayed to simplify the comprehensibility of the figure. Electrode 1 records from the amygdala nuclei (internal contacts) and the anterior part of the middle temporal gyrus (external contacts); electrode 2 records from the hippocampal head (internal contacts) and the middle part of the middle temporal gyrus (external contacts); electrode 3 records from the temporal pole (internal and external contacts); electrode 4 records from the posterior part of the parahippocampal gyrus (internal contacts) and the posterior part of the inferior temporal gyrus; electrode 5 records from the posterior part of the insula (internal contacts) and the anterior part of the superior temporal gyrus (external contact); electrode 6 records from the hippocampal tail (internal contact) and the posterior part of the middle temporal gyrus (external contacts); electrode 7 records from the anterior part of the parahippocampal gyrus (internal contacts) and the anterior part of the inferior temporal gyrus (external contacts); electrode 8 records from the anterior part of the insula (internal contacts) and the middle part of the superior temporal gyrus (external contacts); electrode 9 records from the isthmus of the cingulate gyrus (internal contacts) and the supramarginalis gyrus (external contacts); and electrode 10 records from the orbitofrontal region (internal contacts) and the superior frontal gyrus (external contacts).

Download figure to PowerPoint

Clinical variables

Clinical variables assessed by investigators during the presurgical procedure included the following:

  • 1
    History (history of febrile convulsions or central nervous system (CNS) infections or head injury, age at onset of epilepsy, frequency of secondary generalization),
  • 2
    Initial ictal symptoms (epigastric sensation, other autonomic symptoms, affective and emotional symptoms, dreamy state, sensory illusions or hallucinations),
  • 3
    Early (first half of the seizure) and late (second half of the seizure) ictal signs (oroalimentary automatisms, upper or lower extremity automatisms, nonverbal and verbal automatisms, head and/or eyes deviation and/or tonic posturing, facial expression, vegetative signs). Verbal automatisms were defined as the ictal production of understandable stereotyped words or phrases, for which the patient was subsequently amnesic. Vocalizations (moaning, groaning, howling, and humming) were defined as stereotyped vocal, nonverbal automatisms.
  • 4
    Initial (first 10 s of the seizure) or secondary loss of contact (defined as the absence of nonstereotyped reaction oriented to a visual, motor, verbal, or sensory stimulus)
  • 5
    Seizure duration
  • 6
    Postictal deficits (dysphasia, confusion, and amnesia)

The clinical data presented here were acquired directly by a trained nurse or a medical observer (J.P.V., F.B.), after a standard clinical testing protocol during video-SEEG recordings of seizures. The formal testing protocol included language assessment by asking the patient to describe his or her sensations when present, to name objects from standardized pictures, and if not possible, to point to pictured objects and execute simple commands (close your eyes, stick out your tongue). When the patient did not properly respond, he or she was invited to follow simple commands through imitation (grip the presented hand). If the patient was unable to respond to this, reactivity to visual stimuli was checked (visual field, ocular pursuit). After the seizure, the patient was asked to recall at least one of the presented objects. The testing procedure was carried on until full recovery. Clinical data were retrospectively reviewed on videotape recordings by two independent investigators (J.P.V., L.M.). Seizure onset and termination were determined from the earliest and latest ictal SEEG changes. MRI data were obtained by a 1.5-Tesla device and visually analyzed.

Determination of seizure subtypes

For each patient, the seizures were categorized by one investigator (F.B.) into one of three electrophysiologic subtypes according to the anatomic origin of the initial fast tonic discharge by using visual analysis of SEEG. For each patient, these patterns were reproducible (20,21): medial (M, limbic), lateral (L, neocortical) and medial-lateral (ML).

Patients were classified as the “medial” (M) subtype when the fast tonic discharge (that can be preceded by a preictal periodic spike discharge) was initially observed in the medial structures and when involvement of the temporal neocortex was absent or delayed and took the form of a secondary rhythmic spike discharge (mean delay, 32 s; range, 15–80 s; Fig. 2).


Figure 2. Electrophysiologic example of medial (M) seizure in a patient with medial temporal lobe epilepsy (MTLE) associated with hippocampal sclerosis. A, amygdala; H, hippocampus; EC, entorhinal cortex; T1, superior temporal gyrus; T2, middle temporal gyrus; T3, inferior temporal gyrus; P, parietal region. Seizure starts from the medial region of the temporal lobe (A, H, EC).

Download figure to PowerPoint

As previously described (20,21), patients were classified into the “medial-lateral” (ML) group when the tonic discharge initially affected both the limbic structures and the neocortex (mean delay, 1.55 s; range, 0–3 s). Figure 3A and B shows typical patterns of ML seizures.


Figure 3. Electrophysiologic examples of medial–lateral (ML) seizures in patients with medial temporal lobe dysplastic lesion (A) and anterior lateral and inferior temporal lobe cavernoma (B). Seizures start by a tonic discharge affecting both medial (A, H, EC) and lateral (T1, T2) structures. A, amygdala; H, hippocampus; EC, entorhinal cortex; T1, superior temporal gyrus; T2, middle temporal gyrus; T3, inferior temporal gyrus; P, parietal region.

Download figure to PowerPoint

The “lateral” (L) subtype was defined by an initial tonic discharge involving the neocortex. Propagation to medial structures in the form of a tonic–clonic discharge was absent (66%) or delayed (mean, 20 s; range, 10–30 s; Fig. 4).


Figure 4. Electrophysiologic example of lateral (L) seizure in a patient with a lesion in the first temporal gyrus. This seizure affected exclusively the T1 region without involving the medial structures. A, amygdala; H, hippocampus; TP, temporal pole; T1a, anterior part of the superior temporal gyrus; T1p, posterior part of the superior temporal gyrus; T2, middle temporal gyrus; P, parietal region.

Download figure to PowerPoint

Statistical analysis

The distribution of each clinical variable was compared between subtypes M, L, and ML by using Pearson's χ2 test or Fisher's exact test, according to their condition of validity. For all tests, p values <0.05 (two-sided tests) were considered statistically significant. Statistical analysis was performed by using SPSS software, version 11.1 for Windows.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments

In total, 187 SEEG-recorded seizures were analyzed in 55 patients (3.4 per patient). For each patient, the pattern of coupling between limbic and neocortical structures was reproducible. Patients were then classified into one of the subtypes. Twenty-four patients were identified as M, 13 as L, and 18 as ML.

Medical history, morphologic data, and general characteristics of the population studied

Correlations of medical history, morphologic data, and surgical outcome with electrophysiologic subtypes of TLSs are not discussed in detail in this article, as our aim was to identify electroclinical patterns from SEEG analysis. However, main results are described below. (see also Table 1.)

Table 1. Medical history, magnetic resonance imaging, pathology, and surgery outcome
DataMedial 24 (%)Medial–Lateral 18 (%)Lateral 13 (%)Degree of significance
  1. aSignificant.

History of childhood febrile seizures14 (58.3)4 (22.2)0 p = 0.0004a
Familial history of epilepsy 3 (12.5)4 (22.2)0p = 0.22  
Mean age at onset of epilepsy (yr)7.912.811.2 p = 0.021a
Left temporal lobe epilepsy20 (83.3)13 (72.2)6 (42.6)p = 0.06  
MRI considered normal 7 (29.2)5 (27.7)5 (38.5) 
Hippocampal sclerosis
MRI16 (66.7)4 (22.2)0 p < 0.0001a
Pathology19 (79.2)8 (44.4)0 p < 0.0001a
Medial temporal lesion: MRI04 (22.2)0p = 0.01a
Medial temporal lesion: pathology04 (22.2)0p = 0.01a
Low-grade glioma1 
Focal cortical dysplasia (FCD)1 
Neocortical temporal lesion: MRI2 (8.3)8 (44.4)8 (61.5)p = 0.002a
Neocortical temporal lesion: pathology1 (4.2)6 (33.3)8 (61.5)p = 0.001a
Low-grade glioma000 
Dysembryoplastic neuroepithelial tumor020 
Operated-on patients: 4623 (95.8)15 (83.3) 8 (61.5)p = 0.027a
Resection of the lesion or subarachnoid hemorrhage (SAH): 21 8 (34.8)5 (33.3)8 (100) p = 0.003a
Temporal cortectomy: 2515 (65.2)10 (66.7) 0 
Engel's class I surgical outcome
Global19/23 (82.6)  10/15 (66.6)   6/8 (75)   p = 0.529
Of resection of the lesion or SAH5/8 (62.5)1/5 (20)   6/8 (75)   p = 0.139
Of temporal cortectomy14/15 (93.3)  9/10 (90)     

A history of febrile convulsions was statistically more frequent in M (58.3%) and ML (22.2%) groups compared with the L group (none) (p = 0.0004). Mean age at the onset of the epilepsy was significantly younger for M patients compared with ML and L patients (7.9 in M vs. 12.8 and 11.2 years; p = 0.021). Sixteen patients had right temporal lobe epilepsy (29.1%), and 39 (70.9%) had left temporal lobe epilepsy. Left temporal lobe epilepsy tended to be less frequent in the L group (46.2%) than in the M (83.3%) and ML groups (72.2%), without reaching significance (p = 0.06). This reflected the fact that after phase I investigations, many patients with a clear left lateral temporal epilepsy (in the language-dominant hemisphere) were contraindicated for surgery without further investigation, whereas patients with right lateral temporal epilepsy were selected for SEEG.

Hippocampal sclerosis (HS) on MRI was more frequent in the M group than in the ML and L groups (66.7% in M vs. 22.2% in ML and none in L; p < 0.0001). A medial temporal lesion (associated or not with HS) was specific to the ML subtype (22% in ML vs. 0% in M and L groups; p = 0.01). A neocortical temporal lesion (inferior or lateral) was not specific for the L subtype but was more frequent than in ML and M subtypes (61.5% in L, 44.4% in ML vs. 8.3% in M; p = 0.002).

Forty-six (83.6%) of 55 patients were operated on. No significant difference in Engel's class I surgical outcomes was found between the three groups (82.6% in M vs. 66.6% in ML and 75% in L; p = 0.529). In the M group, 23 (95.8%) of 24 patients were operated on. Of these 23 M patients, 14 (93.3%) of 15 with an anterior temporal cortectomy and five (62.5%) of eight with an amygdalohippocampectomy were Engel class I >2 years after surgery. In the L group, eight (54%) of 13 patients were operated on and six (75%) of eight patients who underwent tailored resection were Engel class I >2 years after surgery. For the operated-on patients of the ML group, the postsurgical outcome was much better for a temporal cortectomy (nine of 10 patients; i.e., 90% of patients in Engel class I) than for resection of the lesion (one of five patients; i.e., 20% of patients in Engel class I). In the ML patients, the temporal cortectomy included any lesion that had been identified, as well as either limbic or neocortical structures, depending on the SEEG definition of the epileptogenic zone.

Semiologic analysis

Initial subjective symptoms

The presence of an initial subjective symptom was reported with approximately the same frequency in the different groups of patients (79%, 83%, and 84% in M, ML, and L groups, respectively; Table 2).

Table 2. Distribution of the main subjective symptoms at seizure onset according to the electrophysiologic subtype
Ictal featuresMedial 24 (%)Medial–Lateral 18 (%)Lateral 13 (%)Degree of significance
  1. aSignificant.

Viscerosensory symptoms19 (79.2)9 (50) 3 (23.1) p = 0.004a
Epigastric sensation11 (45.8)7 (38.9)1 (7.7) p = 0.054
Fear 9 (37.5)4 (22.2)0 p = 0.026a
Dreamy state 7 (29.2)5 (27.8)0p = 0.08 
Auditory hallucination or illusion1 (4.2)2 (11.1)6 (46.2) p = 0.005a
Vertigo2 (8.3)2 (11.1)2 (15.4)p = 0.86 
Visual hallucination or illusion2 (8.3)3 (16.7)4 (30.8)p = 0.21 
Sensory hallucination or illusion (visual, auditory, vestibular) 3 (12.5)6 (33.3)11 (84.6)   p < 0.0001a
Gustatory hallucination2 (8.3)1 (5.6) 0p = 0.78 

Viscerosensory symptoms, including epigastric sensation, thoracic sensation, and warm ascending sensation of arms, were significantly more frequent in M patients (79.2%) than in ML (50%) and especially in L (23.1%) patients (p = 0.004). Among these symptoms, epigastric sensation occurred more frequently in the M and ML subtypes (45.8% and 38.9%) than in the L group (10%), and this tendency almost reached statistical significance (p = 0.054). Fear was reported by none of the patients in the L group and by 37.5% and 22.2% of M and ML patients, respectively (p = 0.026). Dreamy state was not reported in the L group but was experienced by the patients with M and ML seizures in roughly the same proportion (29.2%, 27.8%; p = 0.08).

Among sensory illusions or hallucinations, only the auditory modality significantly differentiated the L from the M and ML groups (46.2% in L vs. 4.2% and 11.1% in M and ML; p = 0.005). Taken together, visual, auditory, and vestibular hallucinations or illusions also were statistically more frequent in L (84.6%) than in M (12.5%) and ML (33.3%) subtypes (p < 0.0001).

Ictal behavior

General ictal characteristics

When considering the whole course of the seizure, the frequency of an ictal loss of contact, irrespective of its latency, did not significantly differ between groups (70.8%, 72.2%, and 69.2% in M, ML, and L groups, respectively; Table 3). In the whole course of the seizures, loss of contact occurred in 74% of left and 62% of right TLS (not statistically significant). Early or late oroalimentary automatisms and upper-limb elementary automatisms were characteristic of M and ML compared with L patients (62.5% and 61.1% in M and ML vs. 15.4% in L; p = 0.014 for oroalimentary automatisms; 66.7% and 55.6% in M and ML vs. 23.1% in L; p = 0.039 for upper-limb elementary automatisms). Verbal automatisms were more frequent in ML compared with M and L patients, regardless of their delay of occurrence (38.9% in ML vs. 16.7% and 0% in M and L; p = 0.03). A nonsignificant trend toward right temporal lateralization of verbal automatisms was noted (31.3% vs. 15.4%), regardless of the delay and subtype of seizure. Vocalization, upper-limb tonic posturing, head and/or eyes deviation did not differ between groups.

Table 3. Distribution of the main ictal and postictal signs, regardless of the timing of their occurrence, according to the electrophysiologic subtype
Ictal and postictal featuresMedial 24 (%)Medial–Lateral 18 (%)Lateral 13 (%)Degree of significance
  1. aSignificant.

Initial or secondary loss of contact17 (70.8)13 (72.2)9 (69.2)p = 1    
Oroalimentary automatisms15 (62.5)11 (61.1)2 (15.4) p = 0.014a
Upper-limb elementary automatisms16 (66.7)10 (55.6)3 (23.1) p = 0.039a
Verbal automatisms 4 (16.7) 7 (38.9)0p = 0.03a
Vocalization 7 (29.2) 7 (38.9)3 (23.1) p = 0.624
Upper-limb tonic posturing 7 (29.2) 5 (27.8)2 (15.4)p = 0.72 
Head and/or eyes deviation 9 (37.5)9 (50) 5 (38.5)p = 0.69 
Long duration (>60 s)23 (95.8)15 (83.3)5 (38.5)  p = 0.0005a
Frequent secondary generalizations (=2/yr) 4 (16.7) 3 (16.7)8 (61.5) p = 0.01a
Postictal confusion6 (25)  7 (38.9)1 (7.7) p = 0.14
Postictal dysphasia12 (50)  11 (61.1)3 (23.1)p = 0.10

A seizure duration of >60 s differentiated M and ML groups from L (95.8% and 83.3% in M and ML vs. 38.5% in L; p = 0.0005). Recurring secondary generalizations (more than two per year) were more frequent in L than in M and ML patients (61.5% in L vs. 16.7% in M and ML; p = 0.01).

Early features

Initial loss of contact (during the first 10 seconds of the seizure) was statistically more frequent in L and in ML than in M groups (53.8% and 38.9% vs. none; p < 0.0001; Table 4). It was encountered in 28% of left and 19% of right temporal seizures (nonstatistically significant) regardless of the seizure subtype. Early oroalimentary automatisms (i.e., occurring during the first half of the seizure) were statistically more frequent in the ML compared with the M and L patients (55.6% vs. 25% and 7.7%; p = 0.015). Likewise, early vocalizations (groaning, howling, moaning, and humming) differentiated ML from the M and L groups (38.9% vs. 12.5% and 7.7%; p = 0.048). Moreover, early verbal automatisms were specific of this subtype (27.8% in ML vs. none in M and L; p = 0.004). Early upper-limb elementary automatisms occurred with the same frequency in patients with M and ML seizures (33.3% and 36.9%) and tended to differ from the L patients (7.7%; p = 0.13). Early upper-limb tonic posturing and early head and/or eyes deviation did not differ between groups.

Table 4. Distribution of the early ictal signs according to the electrophysiologic subtypes
Early ictal featuresMedial 24 (%)Medial–Lateral 18 (%)Lateral 13 (%)Degree of significance
  1. aSignificant.

Initial loss of contact07 (38.9) 7 (53.8)   p < 0.0001a
Early oroalimentary automatisms6 (25)  10 (55.6) 1 (7.7)  p = 0.015a
Early vocalizations (groaning, howling, moaning)3 (12.5)7 (38.9)1 (7.7)  p = 0.048a
Early verbal automatisms05 (27.8)0  p = 0.028a
Early upper-limb elementary automatisms8 (33.3)7 (38.9)1 (7.7)p = 0.13
Early upper-limb tonic posturing2 (8.3) 01 (7.7)p = 0.45
Early head and/or eyes deviation3 (12.5)4 (22.2) 3 (23.1)p = 0.67
Later features

The occurrence of oroalimentary automatisms or upper-limb elementary automatisms during the second half of the seizure was more characteristic of the M than the ML and L patients (58.3% in M vs. 22.2% and 15.4% in ML and L; p = 0.012 for the late oroalimentary automatisms; 58.3% vs. 38.9% and 15.4%; p = 0.039 for the late upper-limb elementary automatisms; Table 5). When occurring later, vocalizations tended to be more frequent in M and L patients (20.8% and 15.4% in M and L vs. none in ML; p = 0.12).

Table 5. Distribution of the late ictal signs according to the electrophysiologic subtypes
Late ictal featuresMedial 24 (%)Medial–Lateral 18 (%)Lateral 13 (%)Degree of significance
  1. aSignificant.

Late oroalimentary automatisms14 (58.3)4 (22.2)2 (15.4)  p = 0.012a
Late upper-limb elementary automatisms14 (58.3)7 (38.9)2 (15.4)  p = 0.039a
Late vocalization 5 (20.8)02 (15.4)p = 0.12
Late verbal automatisms 3 (12.5)2 (11.1)0p = 0.40
Late upper-limb tonic posturing6 (25) 5 (27.8)1 (7.7) p = 0.41
Late head and/or eyes deviation 8 (33.3)6 (33.3)3 (23.1)p = 0.86
Late dysphasia6 (25) 3 (16.7)0p = 0.14
Postictal behavior

The distribution of postictal dysphasia and postictal amnesia was not statistically different between groups (Table 3). Postictal dysphasia was statistically more frequent in left than in right temporal lobe seizures, regardless of the electrophysiologic subtype (61.5% vs. 12.5%; p < 0.001). Among the 26 patients with postictal dysphasia, 24 had left temporal lobe seizures (92%).


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments


In this work, we tried to determine the clinical patterns associated with three electrophysiologically defined models of TLS. These subtypes (M, ML, and L) were characterized according to the anatomic organization of the initial discharge. This distinction was based on SEEG analysis, which allows simultaneous sampling of medial and lateral temporal structures. In previous studies, quantitative analysis of SEEG signals demonstrated that these different electrophysiologic models of TLS can be associated with specific patterns of synchronization between medial and lateral temporal structures (20,21).

Unlike some previous studies, we deliberately chose to include all the patients whose SEEG exploration clearly demonstrated a localized temporal lobe epileptogenic zone, including those who eventually refused to be operated on or in whom surgery was contraindicated, as well as those patients who were not cured by surgery. For the purpose of this study, which was to correlate ictal semiology with the spatiotemporal patterns of the discharge within the temporal lobe, we identified the epileptogenic zone as defined by SEEG, rather than using the spatial extent of the eventual surgical resection. Although with depth studies, a theoretical risk of sampling errors exists, these patients were included in the study because their recordings showed clearly that the epileptic discharge preceded the first clinical sign and because the implantation protocol systematically explored all functional areas of the temporal lobe in their lateral and mesial aspects. The localization of the initial fast discharge can therefore be regarded as the seizure onset area rather than propagation from a distant region.

Significant electroclinical correlations in TLS subtypes

Viscerosensory sensations, epigastric sensation

Many authors have demonstrated the diagnostic value of initial viscerosensory sensations, and especially of epigastric sensation, in differentiating MTLE from LTLE (15,25–27). The present study shows that initial epigastric sensation may help to differentiate the M and the ML from the L subtype. The frequent occurrence of viscerosensory symptoms in the ML subtype is probably due to the limbic component of the network, as suggested by previous stimulation studies (28). This symptom was sometimes, however, reported in pure lateral temporal seizures without early propagation of the neocortical discharge to the hippocampus. This could be explained by an involvement of the insular region, which can produce symptoms similar to those of medial temporal seizures, as previously reported (23,29)

Fear or anxiety

Fear or anxiety also was associated with an epileptogenic network initially involving the medial structures and was found with the same frequency in M and ML seizures. This feature is consistent with the role of limbic structures, particularly the amygdala, in the emotional experience of fear and anxiety as observed during spontaneous seizures and stimulation (1,28,30–32).

Sensory illusions and hallucinations

In this study, as previously proposed (33,34), we separated pure sensory illusions or hallucinations from vivid hallucinations, vivid recollections of scenes, and from the sensation of déjà-vu. The more frequent occurrence of sensory illusion or hallucination (visual, auditory, or vestibular) in LTLS compared with MTLS and MLTLS is consistent with the localizing value of these symptoms (6,7,35,36): Heschl's gyrus for primary auditory hallucinations, more extended and more lateral parts of the superior temporal gyrus (STG) for complex auditory hallucinations (37), temporo-parietal junction involvement for vertigo (38), and basal-temporal gyri and temporo-occipital junction for complex visual hallucinations (39).

Initial loss of contact

An initial loss of contact always reflected an early involvement of the temporal neocortex in a pure L or ML network. When the initial discharge was limited to the limbic structures, the contact was always preserved at the beginning of the discharge, as demonstrated by initial warning or early interactions with the examiner. Staring or behavioral arrest has already been reported as the most common manifestation of neocortical TLS (13) but not as a discriminating feature between M and LTLS (3).

More recently, alteration of consciousness has been found to be more frequent in left and bilateral TLS (40). In this latter study, alteration of consciousness was defined as the alteration of one of its constituent functions, including “the orientation to the examiner, expressive or receptive speech, and postictal memory.” In the present study, loss of contact referred only to the absence of orientation to the examiner and not to speech disturbance, which may explain why it did not have a left lateralization value. This result shows that loss of contact requires extensive involvement of the ipsilateral temporal lobe, and particularly the inferior and lateral structures. We could speculate that ictal disruption of the rich connections between the lateral temporal lobe and association cortices could facilitate inactivation of the normal multisensory integration, which may contribute to loss of contact. Late occurrence of loss of contact in MTLS (with a mean delay of 43.9 s) was frequent and was temporally correlated with a secondary propagation of the discharge to the neocortex. These results are consistent with a recent positron emission tomography (PET) study showing that a larger interictal temporal hypometabolism in MTLE, when associated with loss of contact, was suggestive of a medial–lateral propagation pathway (41).

Seizure duration

Longer seizure duration was characteristic of an early medial discharge (M and ML networks) compared with a pure L discharge. This has not been documented in recent articles comparing MTLS with LTLS (15). This may be related to the technique of depth-electrodes recording used in this study, which allows earlier detection of the discharge onset in MTL structures compared with subdural electrodes (42)

Secondary generalizations

The more frequent occurrence of secondary generalizations in L (60%) compared with M and ML subtypes (10%) has not often been reported in the literature. The lack of significance in previous reports may be due to the great variability of the threshold chosen by the different authors: it ranged from one generalization or more in the whole course of the epilepsy (15) to one generalization or more per month (14). We set the threshold at two generalizations per year on clinical grounds, because it seemed to offer the best discrimination between possible but rare generalizations in MTLS, and recurring generalizations in LTLS.

Oroalimentary automatisms

When considering the whole course of the seizures, oroalimentary automatisms occurred with the same frequency in M and ML seizures. When occurring early during the seizure, they were characteristic of the ML seizure subtype. Their occurrence during the second half of the seizure (i.e., after propagation of the initial discharge) was more characteristic of the M seizure subtype. These results strongly suggest that the emergence of oroalimentary automatisms requires a widespread dysfunction of the medial and temporal neocortical (inferior or lateral) structures. Our results are consistent with previous studies showing that the masticatory automatisms are closely linked to an ictal involvement not only of the amygdala (29,45) but also of the anterior temporopolar region (7,43). They also are consistent with studies showing that electrical stimulation of the amygdala produces masticatory automatisms only when a widespread limbic and temporal neocortical afterdischarge is seen (28,30,44).

Vocal verbal and nonverbal automatisms

The trend toward right temporal lateralization of verbal automatisms is in line with previous studies showing a nondominant temporal lateralization of “normal ictal speech” (45). The main result is that early verbal automatisms were specific of ML patients, regardless of the lateralization of seizures. Vocalizations are frequent in TLS (45). When occurring early in the seizure, they were more frequent in the ML subtype. This is consistent with a previous study on ictal humming, showing that these automatisms occurred when the ictal discharge involved both the limbic and neocortical (superior and middle temporal gyri) structures (46). This suggests that, like verbal automatisms, the occurrence of vocalizations is related to the involvement of a limbic–neocortical temporal network by the ictal discharge.

Upper-limb automatisms

Upper limb automatisms were more frequent during the second phase of M seizures, which was temporally correlated to a more widespread neocortical propagation of the initial limbic discharge. This shows that the types of ictal automatisms and their temporal occurrence are related to the involvement of limbic–neocortical temporal networks by the initial or secondary ictal discharge.

Trends toward other electroclinical correlations

Dreamy state

Even though it did not reach significance, dreamy state was observed only in seizures that initially involved the limbic structures in M and ML networks. Conversely, pure sensory illusions and hallucinations were most specific of LTLS. Recently, experiential auras have been reported to be more frequent in neocortical TLE (15). However, in these latter studies, the term experientialaura included very different symptoms ranging from vague sensory illusions or hallucinations to vivid memory-like recollections or sensation of déjà-vu. Our results reinforce the idea that the experiential phenomena with complex or mnemonic components and those consisting of pure sensory illusions or hallucinations have different anatomic and pathophysiologic substrates. We found that pure sensory hallucinations were related to a neocortical epileptogenic network, whereas dreamy state tended to be related to an epileptogenic network including temporolimbic structures. This is in line with previous works highlighting the role of limbic structures in the dreamy state (47–49) and the role of interactions between medial and lateral structures (34). The apparent discrepancy with the early conclusions of Penfield (2), who stressed the role of neocortical structures in experiential phenomena, has been attributed to the fact that he stimulated mainly the lateral surface of the cortex without recording from the medial part (33).

Nondifferentiating ictal characteristics

In this series, early or late upper-limb tonic posturing, head and eye deviation, late loss of contact, and dysphasia were not specific for any type of TLS, which is consistent with a previous study comparing M and L TLE (15). The left lateralization value of late ictal dysphasia and especially of postictal dysphasia is consistent with previous studies of speech disturbances in TLS (24,45).

Synthesis and perspectives

To summarize, the analysis of 187 SEEG-recorded TLS in 55 patients shows that the M, ML, and L localizations of the initial discharge determine distinct sequences of ictal semiology.

The M subtype (limited to the limbic structures) is characterized by initial epigastric sensation or viscerosensitive symptoms, fear, dreamy state, longer seizure duration (>1 min); notably a delayed loss of contact and delayed oroalimentary and upper-limb automatisms.

The L subtype is characterized by an initial sensory illusion or hallucination (mainly auditory), an initial loss of contact (when present), a shorter duration of seizures and frequent secondary generalizations.

The ML subtype shares the same initial subjective symptoms as M and also is characterized by long seizure duration. Unlike pure M seizures, ML seizures are characterized by an earlier loss of contact and earlier oroalimentary, verbal, and vocal automatisms. In TLE, the existence of seizures related to a widespread temporal epileptogenic zone encompassing both limbic and neocortical structures has long been recognized (1,7). Recently, EEG quantification studies demonstrated a strong functional coupling between these two structures in this type of seizure (21), described as ML. However, MLTLS have not previously been described as a separate entity. The present study for the first time individually characterizes ML seizures in terms of their electroclinical correlations.

The association of MTLS with a younger age at onset, a history of childhood febrile seizures, and HS is consistent with the previously well-described MTLE syndrome (8,9,11). Likewise, the association of LTLS with an older age at onset and a neocortical temporal lesion is keeping with the recognized syndrome of LTLE (13–16). In our study, medial temporal lesions (with or without HS) were associated with ML seizures. However, the identification of ML seizures, which also can be encountered in patients with HS or neocortical temporal lesions, demonstrates the concept that the epileptogenic zone may be more widespread than the lesion and encompass both limbic and neocortical areas within the temporal lobe.

The goal of our study was not to assess correlations between the seizure subtypes and eventual surgical outcome, because this retrospective analysis did not influence the surgical procedure. However, it seems highly likely that such data could have a potential major influence on current surgical practice, as suggested by the much better outcome of temporal cortectomy compared with resection limited to the lesional and perilesional area in ML patients.

In these cases, the presence of ML semiologic characteristics could assist the surgeon in achieving optimal delineation of the volume of tissue to be excised, in other words, choosing between a resection limited to the lesion or a larger temporal cortectomy. It could also indicate further presurgical investigations to determine the respective roles of limbic and neocortical regions in the initial organization of seizures.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments

Acknowledgment:  We thank Pr. J. Régis and Pr. J.M. Scarabin for stereotactic placement of the electrodes. We thank Dr. Roch Giorgi for statistical analysis.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  6. Acknowledgments
  • 1
    Bancaud J, Talairach J, Bonis A, et al. La stéréoélectroencéphalographie dans l'épilepsie: informations neurophysiopathologiques apportées par l’investigation fonctionnelle stereotaxique. Paris : Masson, 1965.
  • 2
    Penfield W, Jasper H. Epilepsy and the functional anatomy of the human brain. Boston : Little, Brown, 1954.
  • 3
    Escueta AV, Bacsal FE, Treiman DM. Complex partial seizures on closed-circuit television and EEG: a study of 691 attacks in 79 patients. Ann Neurol 1982;11: 292300.
  • 4
    Walsh G, Delgado-Escueta A. Type II complex partial seizures: poor results of anterior temporal lobectomy. Neurology 1984;34: 113.
  • 5
    Williamson PD, Spencer SS. Clinical and EEG features of complex partial seizures of extratemporal origin. Epilepsia 1986;27(suppl 2):S4663.
  • 6
    Bancaud J. Epileptic attacks of temporal lobe origin in man. Jpn J EEG EMG 1981; (suppl):6171.
  • 7
    Wieser H. Electroclinical features of the psychomotor seizures. London : Butterworths, 1983.
  • 8
    Commission on classification and terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30: 38999.
  • 9
    French J, Williamson P, Thadani V, et al. Characteristics of medial temporal lobe epilepsy, I: results of history and physical examination. Ann Neurol 1993;34: 77480.
  • 10
    Hajek M, Antonini A, Leenders KL, et al. Mesiobasal versus lateral temporal lobe epilepsy: metabolic differences in the temporal lobe shown by interictal 18F-FDG positron emission tomography. Neurology 1993;43: 7986.
  • 11
    Williamson P, French J, Thadani V, et al. Characteristics of medial temporal lobe epilepsy, II: interictal and ictal scalp electroencephalography, neuropsychological testing, neuroimaging, surgical results and pathology. Ann Neurol 1993;34: 7817.
  • 12
    Ebersole JS, Pacia SV. Localization of temporal lobe foci by ictal EEG patterns. Epilepsia 1996;37: 38699.
  • 13
    Pacia SV, Devinsky O, Perrine K, et al. Clinical features of neocortical temporal lobe epilepsy. Ann Neurol 1996;40: 72430.
  • 14
    Burgerman R, Sperling M, French J, et al. Comparison of medial versus neocortical onset temporal lobe seizures: neurodiagnostic findings and surgical outcome. Epilepsia 1995;36: 66270.
  • 15
    Gil-Nagel A, Risinger M. Ictal semiology in hippocampal versus extrahippocampal temporal lobe epilepsy. Brain 1997;120: 18392.
  • 16
    Williamson P, Engel P, Munari C. Anatomic classification of localization-related epilepsies. In: EngelJ, PedleyT, eds. Epilepsy: a comprehensive textbook. New York : Lippincott-Raven, 1998: 240516
  • 17
    Schramm J, Kral T, Grunwald T, et al. Surgical treatment for neocortical temporal lobe epilepsy: clinical and surgical aspects and seizure outcome. J Neurosurg 2001;94: 3342.
  • 18
    Walczak TS. Neocortical temporal lobe epilepsy: characterizing the syndrome. Epilepsia 1995;36: 6335.
  • 19
    Munari C, Tassi L, Kahane P, et al. Analysis of clinical symptomatology during stereo-EEG recorded mesiotemporal lobe seizures. In: WolfPW, ed. Epileptic seizures and syndromes. London : John Libbey, 1994: 33558.
  • 20
    Bartolomei F, Wendling F, Vignal J, et al. Seizures of temporal lobe epilepsy: identification of subtypes by coherence analysis using stereo-electro-encephalography. Clin Neurophysiol 1999;110: 174154.
  • 21
    Bartolomei F, Wendling F, Bellanger J, et al. Neural networks involved in temporal lobe seizures: a nonlinear regression analysis of SEEG signals interdependencies. Clin Neurophysiol 2001;112: 174660.
  • 22
    Talairach J, Bancaud J, Szickla G, et al. Approche nouvelle de la chirurgie de l'épilepsie: methodologie stéréotaxique et résultats thérapeutiques. Neurochirurgie 1974;20(suppl 1):1240.
  • 23
    Isnard J, Guenot M, Ostrowsky K, et al. The role of the insular cortex in temporal lobe epilepsy. Ann Neurol 2000;48: 61423.
  • 24
    Serafetinides EA, Falconer MA. Speech disturbances in temporal lobe seizures: a study in 100 epileptic patients submitted to anterior temporal lobectomy. Brain 1963;86: 33346.
  • 25
    Henkel A, Noachtar S, Pfander M, et al. The localizing value of the abdominal aura and its evolution: a study in focal epilepsies. Neurology 2002;58: 2716.
  • 26
    Pfander M, Arnold S, Henkel A, et al. Clinical features and EEG findings differentiating mesial from neocortical temporal lobe epilepsy. Epileptic Disord 2002;4: 18995.
  • 27
    Williamson PD, Thadani VM, French JA, et al. Medial temporal lobe epilepsy: videotape analysis of objective clinical seizure characteristics. Epilepsia 1998;39: 11828.
  • 28
    Fish DR, Gloor P, Quesney FL, et al. Clinical responses to electrical brain stimulation of the temporal and frontal lobes in patients with epilepsy. Pathophysiological implications. Brain 1993;116: 397414.
  • 29
    Penfield W, Faulk ME Jr. The insula; further observations on its function. Brain 1955;78: 44570.
  • 30
    Gloor P. The temporal lobe and limbic system. New York : Oxford University Press, 1997.
  • 31
    Wieser HG. Mesial temporal lobe epilepsy versus amygdalar epilepsy: late seizure recurrence after initially successful amygdalotomy and regained seizure control following hippocampectomy. Epileptic Disord 2000;2: 14152.
  • 32
    Biraben A, Taussig D, Thomas P, et al. Fear as the main feature of epileptic seizures. J Neurol Neurosurg Psychiatry 2001;70: 18691.
  • 33
    Gloor P. Experiential phenomena of temporal lobe epilepsy: facts and hypotheses. Brain 1990;113: 167394.
  • 34
    Bancaud J, Brunet-Bourgin F, Chauvel P, et al. Anatomical origin of deja vu and vivid “memories” in human temporal lobe epilepsy. Brain 1994;117: 7190.
  • 35
    Bancaud J. Clinical symptomatology of epileptic seizures of temporal origin. Rev Neurol ( Paris ) 1987;143: 392400.
  • 36
    Wieser H, Engel JJ, Williamson P, et al. Surgically remediable temporal lobes syndromes. In: EngelJJr, ed. Surgical treatment of the epilepsies, 2nd ed. New York : Raven Press, 1993: 4963.
  • 37
    De Graaf JB, Liegeois-Chauvel C, Vignal JP, et al. Electrical stimulation of the auditory cortex. In: LüdersH, NoachtarS, eds. Epileptic seizures: pathophysiology and clinical semiology. New York : Churchill Livingston, 2000.
  • 38
    Kahane P, Hoffmann D, Minotti L, et al. Reappraisal of the human vestibular cortex by cortical electrical stimulation study. Ann Neurol 2003;54: 61524.
  • 39
    Bancaud J, Talairach J. Crises du lobe temporal. Paris : Sanofi, 1993.
  • 40
    Lux S, Kurthen M, Helmstaedter C, Hartje W, et al. The localizing value of ictal consciousness and its constituent functions: a video-EEG study in patients with focal epilepsy. Brain 2002;125: 26918.
  • 41
    Chassoux F, Semah F, Bouilleret V, et al. Metabolic changes and electro-clinical patterns in mesio-temporal lobe epilepsy: a correlative study. Brain 2004;127: 16474.
  • 42
    Spencer SS, Spencer DD, Williamson PD, et al. Combined depth and subdural electrode investigation in uncontrolled epilepsy. Neurology 1990;40: 749.
  • 43
    Munari C, Bancaud J, Bonis A, et al. Role of the amygdala in the occurrence of oro-alimentary signs of during epileptic seizures in man (author's transl). Rev Electroencephalogr Neurophysiol Clin 1979;9: 23640.
  • 44
    Bancaud J, Talairach J, Morel P, et al. Ammon's horn and amygdaline nucleus: clinical and electric effects of their stimulation in man. Rev Neurol ( Paris )1966;115: 32952.
  • 45
    Gabr M, Luders H, Dinner D, et al. Speech manifestations in lateralization of temporal lobe seizures. Ann Neurol 1989;25: 827.
  • 46
    Bartolomei F, Wendling F, Vignal JP, et al. Neural networks underlying epileptic humming. Epilepsia 2002;43: 100112.
  • 47
    Halgren E, Walter RD, Cherlow DG, et al. Mental phenomena evoked by electrical stimulation of the human hippocampal formation and amygdala. Brain 1978;101: 83117.
  • 48
    Gloor P, Olivier A, Quesney LF, et al. The role of the limbic system in experiential phenomena of temporal lobe epilepsy. Ann Neurol 1982;12: 12944.
  • 49
    Bartolomei F, Barbeau E, Gavaret M, et al. Role of the rhinal cortices in déjà vu-déjà vécu and reminiscence of memories: a study of electrical stimulation in patients with epilepsy. Neurology 2004;63: 85864.