Ictal Semiology Important for Electrode Implantation and Interpretation of Stereoelectroencephalography

Abstract

Scalp video-electroencephalography (video-EEG) monitoring should be analyzed thoroughly to preoperatively evaluate stereoelectroencephalography (SEEG). Formulating the working hypotheses for the epileptogenic zone (EZ) considering “anatomo-electroclinical correlations” is the most crucial step, which determines the placement of SEEG electrodes. If these hypotheses are insufficient, precise EZ identification may not be achieved during SEEG recording.

In ictal semiology analysis, temporal and spatial patterns with reference to ictal EEG changes are emphasized. In frontal lobe epilepsy, seizures often begin with relatively widespread synchronous activity, and complex motor symptoms manifest within seconds. Due to the wide area involved and intense interhemispheric connectivity, a comprehensive evaluation is often required. Hypotheses are formulated on the basis of the motor symptoms and emotional manifestations that are related to the prefrontal cortices. In temporal lobe epilepsy, EEG onset often precedes clinical onset. Propagation from the EZ to locations within and outside of the temporal lobe is examined from both the EEG and semiological standpoint. The characteristics of contralateral versive seizures, contralateral tonic seizures, and frequent focal onset bilateral tonic-clonic seizures indicate a higher risk of temporo-perisylvian epilepsy. In parietal/occipital lobe epilepsy, despite that some symptoms result from activity in the immediate vicinity, stronger connectivity with other regions usually contributes to the generation of prominent ictal semiology. Hence, multilobar electrode placement is often useful in practice. For insular epilepsy, it is important to understand the anatomy, function, and networks between other regions. A semiological approach is one of the most important clues for electrode implantation and interpretation of SEEG.

Introduction

For over 50 years, intracranial implanted electroencephalography has been utilized for the presurgical evaluation of patients with drug-resistant focal epilepsy. Following the classical methodology, refined stereoelectroencephalography (SEEG) has been reintroduced in Japan in recent years, which involves the evaluation of brain networks primarily associated with epileptic seizures, such as the epileptogenic zone (EZ), encompassing the epilepsy focus (an abnormal brain region where epileptic seizures occur) and epileptogenic legion (a lesion associated with epileptic seizures). It is expected that the application of SEEG will significantly expand because of its utility in complicated patients for whom traditional epilepsy surgery is less suitable, including those with negative MRI findings, also known as nonlesional cases, or those with probable EZ located in deep brain areas, such as focal cortical malformations in the white matter, operculum, insula, and mesial or inferior cerebral cortices. Even in major epilepsy centers in North America, where presurgical evaluations were used to be performed primarily with subdural grid electrodes (SDG), there is a paradigm shift toward adopting refined SEEG. It is believed that this change stems from the ability of SEEG to determine the EZ more accurately than SDG in certain patients. Nevertheless, notably, SEEG is not a three-dimensional counterpart of SDG, and its presurgical evaluation, particularly the hypothesis formulation regarding the EZ, electrode implantation planning, and interpretation of SEEG, differs from that of SDG to some extent.¹⁾ Performing SEEG necessitates a thorough presurgical evaluation, mainly the analysis of ictal semiology and scalp electroencephalography (EEG). This unique interpretation should be understood during SEEG electrode implantation. Inadequate presurgical assessment or understanding when conducting SEEG can lead to inappropriate conclusions and potentially hinder favorable postsurgical seizure outcomes. This section provides an overview of the important clinical approaches that are required for SEEG. Notably, this section is based on the content presented at the 46th Annual Meeting of the Epilepsy Surgery Society of Japan held on January 27, 2023. There was no requirement for consent because this study involved no human subjects.

History and Current State of SEEG

SEEG was initially devised by Professors Jean Talairach and Jean Bancaud in France in the 1950s and 1960s. Since then, it has been in use and developed for more than 50 years. SEEG involves the placement of depth electrodes in multiple brain regions in a three-dimensional manner with the use of a stereotactic system. Talairach coordinates, based on the anterior and posterior commissures, are used as a standardized anatomical positioning reference system among patients. In major epilepsy centers in North America, a significant paradigm shift over the past 5-10 years has been observed, transitioning from SDG to refined SEEG as the presurgical evaluation method for epilepsy surgery.^2-4) The increasing number of patients with extratemporal epilepsy and nonlesional epilepsy, as well as the decreasing proportion of mesial temporal lobe epilepsy, which used to be the primary focus of epilepsy surgery, are considered to be associated with the increasing use of presurgical evaluation with SEEG.⁵⁾ Despite that some related equipment remains unapproved, such as anchor bolts, SEEG has been introduced in Japan, and there has been an increasing number of patients in various institutes. The purpose of SEEG, which is to evaluate networks related to epileptic seizures, is fundamentally different from the notion of SDG put on the surface of the cerebral cortex or standard depth electrodes placed bilaterally in the temporal lobes to assess lateralization.^6,7) Although practices may vary among institutes, Gonzalez-Martinez et al. were able to illustrate one approach for selecting intracranial electrodes for presurgical evaluation.⁸⁾

Principles and Theoretical Background of SEEG

The principles of SEEG are based on “anatomo-electroclinical correlations” to verify working hypotheses for epilepsy surgery mainly through the correlation of ictal semiology and other noninvasive evaluations. In presurgical evaluation via SEEG, the formulation of working hypotheses for electrode placement locations based on the “anatomo-electroclinical correlations” related to epileptic seizures in each patient is greatly important. Identification of the EZ during invasive SEEG evaluation could be compromised if these working hypotheses are insufficient or inaccurate. Presurgical evaluation using SEEG can be divided into three steps:

Step 1: Noninvasive presurgical evaluation

Step 2: Hypothesis formulation for the EZ on the basis of the notion of “anatomo-electroclinical correlations”

Step 3: Evaluation during SEEG implantation

Of these steps, Step 1 provides the essential clinical approach required for practicing SEEG. This involves preparing as much information as possible via multimodality noninvasive evaluation and establishment of accurate working hypotheses for the EZ based on “anatomo-electroclinical correlations.” The precise “anatomo-electroclinical correlation” hypotheses must be generated in the multidisciplinary patient management conference using the results of various noninvasive evaluations. Within this context, ictal semiology recorded in high-definition videos during long-term video-EEG monitoring, along with scalp EEG findings (interictal and ictal periods), is primarily important. Moreover, 3T-MRI of the brain, magnetoencephalography (MEG), fluorodeoxyglucose-positron emission tomography, and subtraction of ictal-interictal SPECT coregistered to MRI (SISCOM) are utilized to visualize the differences between ictal and interictal states by overlaying these on MRI. Furthermore, sufficient knowledge of the brain networks involved in the progression of seizure semiology (early and late spread) is necessary. MEG is useful for determining irritable zones within deep brain regions, and SISCOM visualizes epilepsy networks involving deep brain areas; both are valuable tools for preoperative SEEG assessment. A detailed analysis of seizure semiology concerning “anatomo-clinical correlates,” including various deep structures such as the insular, operculum, and cingulate cortex, forms the basis of the implantation strategy for SEEG. Furthermore, in some institutes, in nonlesional cases of drug-resistant focal epilepsy, various methods such as 7T-MRI, voxel-based morphometry, and simultaneous EEG-functional MRI (fMRI) recording have been employed, besides 3T-MRI. The working hypotheses for the EZ in the presurgical phase (Step 2) are derived after a thorough analysis and consideration of the findings from these noninvasive evaluations and the patterns of ictal semiology in terms of temporal and spatial propagations. The clinical characteristics of each epilepsy syndrome are discussed in the following sections.

Frontal Lobe Epilepsy (FLE)

FLE is the second most common type of focal epilepsy after temporal lobe epilepsy (TLE). The frontal lobe is extensive and has strong connectivity within and between hemispheres via cortico-subcortical and cortico-cortical connections. These characteristics contribute to seizures in FLE, which spread rapidly within related regions, requiring a rather thorough evaluation with SEEG, which frequently includes the parietal, anterior temporal, and insular areas. Without SEEG, it can be difficult to accurately assess FLE, especially in nonlesional cases.

Motor semiology is a key element of FLE. Motor symptoms of and areas that are related to frontal lobe seizures (FLE and epilepsy arising from other cortical areas and manifesting as FLE in terms of ictal semiology) have been thoroughly explored using SEEG. Bonini et al. explored SEEG and ictal semiology in 54 patients with FLE and grouped the features of semiology and relevant brain regions into four categories⁹⁾ (Fig. 1). Group 1 involved the precentral/premotor regions, associated with elementary motor signs such as clonic seizures, tonic/dystonic seizures, and head/eye versions. Group 2 encompassed premotor/prefrontal regions, associated with nonintegrated gestural motor behaviors including proximal stereotypies, hyperkinetic seizures, and chapeau movements reminiscent of the face of “Kabuki” actors. Group 3 involved the anterior prefrontal region, contributing to natural movements (reaching, grasping, pedaling, kicking, tapping, rocking, hitting, etc.) and distal stereotypies (repetitive behaviors) with mild changes in facial expression. Group 4 included the ventromedial prefrontal cortex and anterior temporal areas, revealing changes in emotional expression such as fearful behavior and agitation. The semiology and anatomical locations of Groups 1-3 revealed the most important finding: Semiological characteristics were organized in a rostral-caudal manner, with more complicated and highly integrated motor symptoms associated with more rostral frontal cortices. The relationships between semiology and anatomical locations do not match completely, as demonstrated with the premotor region, which may exhibit elementary motor signs as well as more complex movements, whereas seizures arising from the prefrontal cortex lack elementary motor signs. Additionally, such ictal semiology may arise from extrafrontal regions, especially the temporal lobe, anterior insula, and parietal lobe.

Fig. 1.

Four groups of frontal lobe epilepsy patients based on ictal semiology and relevant anatomical locations evaluated with SEEG.

Group 1 was distinguished by elementary motor signs and involvement of the precentral and premotor regions. Group 2 was characterized by elementary motor signs and nonintegrated gestural motor behavior and involved the premotor and prefrontal regions. Group 3 was described by integrated gestural motor behavior and involved the anterior prefrontal region. Group 4 was characterized by seizures with fearful behavior and involved the ventromedial prefrontal cortex and the anterior temporal structures. The semiological characteristics were organized in a rostral–caudal manner.

Singh et al. analyzed SEEG and emotional ictal semiology, particularly of prefrontal seizures that are associated with emotions, in 42 prefrontal epilepsy patients, in detail¹⁰⁾ (Fig. 2). They identified four groups using cluster and principal component analysis: Group 1, joviality/prosocial behavior involving feelings of happiness, singing, rhythmic movements, and social conversation, which is believed to propagate from the frontopolar to the orbitofrontal and pre-SMA regions; Group 2, passive fear characterized by astonishment, freezing, and hiding of the face, which is believed to propagate from the posterior orbitofrontal and amygdala to the anterior cingulate cortex; Group 3, active threat response that involves expressions of astonishment, shouting, hitting, kicking, and movement, which likely propagated from the ventromedial prefrontal cortex (anterior cingulate cortex) to the amygdala and temporal pole; and Group 4, aggression, including growling and hitting, which suggests involvement of the dorsolateral prefrontal cortex.

Fig. 2.

Four groups of patients with prefrontal seizures based on emotional ictal semiology and relevant anatomical locations and seizure network evaluated with SEEG.

Group 1 was described by joviality/prosocial behavior with feelings of happiness, singing, rhythmic movements, and social conversations, which were probably correlated with propagation from the frontopolar to the orbitofrontal and pre-SMA regions. Group 2 was characterized by passive fear including astonishment, freezing, and hiding the face, which probably correlated with propagation from the posterior orbitofrontal and amygdala to the anterior cingulate cortex. Group 3 was distinguished by an active threat response involving expressions of astonishment, shouting, hitting, kicking, and movement, which probably correlated with propagation from the ventromedial prefrontal cortex (anterior cingulate cortex) to the amygdala and temporal pole. Group 4 was characterized by aggression, including growling and hitting, involving the dorsolateral prefrontal cortex.

Temporal Lobe Epilepsy (TLE)

TLE is the most common type of focal epilepsy. The temporal lobe is highly networked within the temporal lobe and with the surrounding areas;¹¹⁾ therefore, seizure propagation within and outside the temporal lobe should be considered. In practical SEEG electrode placement, the mesial-lateral temporal regions are commonly covered with an orthogonal electrode implantation. This approach allows for the consideration of more localized treatments including laser interstitial thermal therapy or resection with hippocampal preservation.

TLE is not a single entity and is generally categorized into mesial TLE (MTLE), neocortical (lateral) TLE, temporo-polar epilepsy, temporal-plus epilepsy, and bilateral TLE.¹²⁾ MTLE is characterized by frequent auras (epigastric sensation or viscerosensory aura, fear, etc.), automatisms, relatively long seizure duration, and infrequent generalized tonic-clonic seizures. Neocortical TLE may have auditory, visual, or vertiginous auras, and automatisms are rare. The seizure duration tends to be shorter, whereas the frequency of secondary generalization is higher. Due to the high excitability of the hippocampus, neocortical TLE sometimes generates the secondary MTLE, which can cause mixed features of MTLE and neocortical TLE. Temporo-polar epilepsy involves connectivity between the limbic system and inferior/basal frontal areas via the uncinate fasciculus, which often shows frontal lobe epilepsy-like symptoms (frontal lobe seizures). The rapid appearance of symptoms as soon as ictal EEG changes occur is a distinctive feature of this condition. Temporal-plus epilepsy implies TLE electroclinically but may involve or originate outside of the temporal lobe and is further subdivided into the following three types: temporo-perisylvian epilepsy, temporo-frontal epilepsy, and temporo-parieto-occipital epilepsy^13,14) (Fig. 3). Temporo-perisylvian epilepsy is distinguished by sensations of constriction in the throat and larynx, dyspnea, abnormal sensations in the mouth and face, shivering, piloerection, abnormal pain perception, and gustatory hallucinations. Temporo-frontal (orbitofrontal) epilepsy tends to cluster during sleep; exhibits brief postictal confusion; and features autonomic symptoms, urgency of urination, motor arrest, altered consciousness, and hyperkinetic seizures. Temporo-parieto-occipital epilepsy is defined by illusions and hallucinations of auditory, visual, and vertiginous sensations.

Fig. 3.

Percentage of patients in whom each particular delineated region is part of the epileptogenic zone in temporal-plus epilepsy patients.

The epileptogenic zones of patients with temporal-plus epilepsy include the orbitofrontal (violet), suprasylvian operculum (brown), insula (red), and temporo-parieto-occipital junction (green).

Parietal Lobe Epilepsy (PLE)

The parietal lobe can be identified by strong connectivity with other cortical regions and is often regarded as the “great imitator.” The precuneus shows functional connectivity with various subregions of the sensorimotor, cognitive, limbic, and visual systems.¹⁵⁾ The fronto-temporo-parietal network also demonstrates significant connectivity with the frontal and temporal lobes. SEEG often involves electrode placement in multiple lobes, including bilateral implantation.

Ictal semiology from the superior parietal lobule may involve sensations of vertigo, emotional changes such as fear, and versive seizures to the same side (propagated to the inferior parietal lobule).¹⁶⁾ Seizures that originate from the inferior parietal lobule, that is, the supramarginal and angular gyri, can generate changes in body image (especially on the nondominant side) and emotional changes such as fear. Seizures originating from the parietal operculum may manifest as somatosensory symptoms, such as tingling, numbness, pain, and thermal sensations. Ictal semiology, which results from propagation to other associated areas in PLE, includes clonic, versive, tonic, and hyperkinetic seizures on the contralateral side if the seizure activity propagates to the frontal lobe. Propagation to the anterior and mesial temporal lobe may lead to automatic behaviors, whereas propagation to the posterior and lateral parts of the occipital and temporal lobes may result in visual illusions, hallucinations, complex visual phenomena, and auditory symptoms.

Posterior Cingulate Epilepsy

Posterior cingulate epilepsy exhibits variable symptoms and propagation patterns.¹⁷⁾ Seizure onset often involves minor alterations in consciousness or motor arrest, as shown in temporal lobe seizures.¹⁸⁾ Propagation to the mesial temporal lobe or inferior parietal lobule via the attention-recognition network can result in feelings of familiarity and deja vu as focal impaired awareness seizures (FIAS). Motor symptoms may result from propagation to the frontal, parietal, and anterior cingulate cortices. Feelings of unfamiliarity as focal aware seizures may arise from propagation to the temporal lobe.

Occipital Lobe Epilepsy

Similar to the parietal lobe, the occipital lobe has strong connectivity with other cortical regions. The dorsal network involves seizures that result from the area above the calcarine fissure, which has connectivity with frontal and parietal lobes via the inferior fronto-occipital fasciculus. The ventral network involves seizures that occur from the area below the calcarine fissure, which is connected to the temporal lobe through the inferior longitudinal fasciculus. SEEG often involves electrode placement in multiple lobes, including bilateral implantation. Noninvasive visual functional assessments including visual evoked potential/field, diffusion tensor imaging, and fMRI are essential for preoperative evaluation. During seizures, visual symptoms may include simple or complex visual hallucinations, illusions, blindness, and visual perseveration.¹⁹⁾ Nonvisual symptoms may involve symptoms that are related to eye movement, including eye deviation, eye clonic seizures, nystagmus, repetitive blinks, and eye fluttering. Additionally, symptoms may arise in the regions where the propagation takes place.

Insular Epilepsy

In insular epilepsy, the connectivity with the surrounding lobes, including the frontal, temporal, and parietal lobes must be considered;²⁰⁾ hence, the secondary hypothesis of EZ includes FLE, TLE, and PLE.²¹⁾ Seizure activity often propagates to the contralateral side, which leads to frequent bilateral SEEG electrode placement. The analysis of ictal semiology regarding the anatomy, function, and networks of the insular cortex is essential. Specifically, seizures that originate from the anterior insula may result in motor symptoms such as hyperkinetic seizures, asymmetric tonic seizures, and elementary facial contractions, as well as autonomic symptoms, fear, and nonmotor FIAS²²⁾ (Fig. 4). Seizures that originate from the posterior insula may involve abnormal sensations such as tingling, electrical current sensations, warmth, or painful sensations and even premonitory visceral sensations such as abdominal, epigastric, or chest sensations.

Fig. 4.

Spread patterns of insula seizures to symptomatogenic zones.

Insular seizures include perisylvian, frontal, and temporal symptoms and even epileptic spasms, suggesting that insular epilepsy can exhibit various ictal semiology depending on the spread patterns.

Conclusions

In the hypothesis formulation for EZ in the presurgical evaluation prior to implanting SEEG, noninvasive preoperative assessments, particularly detailed long-term video-EEG analysis based on “anatomo-electroclinical correlations,” are important. Analyzing ictal semiology, including subtle changes in expression, and considering which regions or networks might be involved in seizures, are crucial. Understanding the anatomy, function, and network (seizure propagation) of each epilepsy syndrome is also important. The ictal semiology for each epilepsy syndrome outlined in this section may not necessarily be applicable to all patients. Nonetheless, it is generally important for the hypothesis formulation of EZ with SEEG and electrode placement planning. Subsequently, this data collection may be more applicable to artificial intelligence analysis and clinical derivation and validation. Ultimately, SEEG is employed to validate these hypotheses for the EZ; formulation of EZ hypothesis prior to SEEG is most crucial.

Abbreviations

artificial intelligence: AI, diffusion tensor imaging: DTI, electroencephalography: EEG, epileptogenic zone: EZ, fluorodeoxyglucose positron emission tomography: FDG-PET, focal aware seizures: FAS, focal impaired awareness seizures: FIAS, frontal lobe epilepsy: FLE, functional MRI: fMRI, generalized tonic-clonic seizures: GTCS, laser interstitial thermal therapy: LITT, magnetoencephalography: MEG, mesial temporal lobe epilepsy: MTLE, occipital lobe epilepsy: OLE, parietal lobe epilepsy: PLE, stereoelectroencephalography: SEEG, subdural grid electrodes: SDG, subtraction of ictal-interictal SPECT co-registered to MRI: SISCOM, temporal lobe epilepsy: TLE, visual evoked field: VEF, visual evoked potential: VEP, voxel-based morphometry: VBM

Conflicts of Interest Disclosure

All authors declare no potential financial interests relating to this manuscript.

Department of Epilepsy, Movement Disorders and Physiology (Kyoto University Graduate School of Medicine) is the Industry-Academic Collaboration Courses, supported by Eisai Co., Ltd., NIHON KOHDEN CORPORATION, Otsuka Pharmaceutical Co., and UCB Japan Co., Ltd.