Presurgical Evaluation for Epilepsy Including Intracranial Electrodes

Published on 13/03/2015 by admin

Filed under Neurosurgery

Last modified 22/04/2025

Print this page

This article have been viewed 1532 times

Chapter 108 Presurgical Evaluation for Epilepsy Including Intracranial Electrodes

Ana Luisa Velasco, Francisco Velasco, Bernardo Boleaga, José María Núñez, David Trejo

Medically Intractable Epilepsy

As many as 25% to 30% of epileptic patients fail to respond to adequate antiepileptic treatment.¹ Medically intractable (or refractory) epilepsy is commonly conceived as that which occurs when satisfactory seizure control cannot be achieved with any of the potentially available effective antiepileptic drugs (AEDs), alone or in combination, at doses or levels not associated with unacceptable side effects.² On the other hand, surgery has a success rate of about 75% (depending on the site of surgery, the type of surgery performed, and the underlying etiology).^3,⁴ In spite of this, surgical treatment of medically intractable epilepsy is often delayed or withheld. Referral for epilepsy surgery may take 20 to 25 years, resulting in a number of avoidable seizure-related deaths, including drowning, motor vehicle accident, fatal status epilepticus, and sudden unexpected death in eplilepsy (SUDEP).⁵ In children, appropriate timing for surgical procedures is critical, as lack of seizure control may interfere with consolidation of cognitive and motor functions in the developing brain.⁶

Other causes of poor response to medical treatment are noncompliance and presence of pseudoepileptic seizures alone or combined with epileptic ones, as well as incorrect classification of seizures and therefore incorrect pharmacologic treatment.⁷ Therefore, the first step in evaluating surgical candidates is to confirm that the proper diagnosis has been identified and to determine that appropriate medical treatment has been utilized.

Patients’ Perceptions

Patients’ satisfaction with achieved medical treatment results often depends on professional and social circumstances, as well as the type of seizures and the time of day in which they occur.⁷ Some patients seek surgical treatment for reasons other than medical intractability: intolerable side effects of AEDs, women who desire to get pregnant but are concerned about potential teratogenic medication effects, and avoidance of the social stigmata of the disease are some of the reasons patients express in requesting surgery. Self-assessment of quality of life is an important step in evaluating the patient’s perception of their disease. At our clinic, we use the questionnaires QOLIE-31 for adults^8,⁹ and QOLIE-48-AD for adolescents.^10,¹¹

Epilepsy surgery has three main goals: eliminate or decrease epileptic seizures, prevent neurologic deficit due to surgery, and improve the quality of life. In order to achieve these goals, a multidisciplinary team is required to solve the surgical questions: Where do the patient’s seizures start? Are there functional or eloquent areas involved? What is the prognosis after surgery? Answering these questions allows decisions regarding what type of surgery is indicated and how to customize the procedure to meet the patient’s needs. Currently, there are many diagnostic studies that can be performed, but it is important to remember that there is no single test that can be considered the gold standard for making a diagnosis or predicting the outcome. It is the clinician’s ability to correctly interpret and assess the quality and concordance between studies that usually leads to a correct diagnosis and successful surgery.¹² The diagnostic workup can be divided into noninvasive and invasive phases.

Phase 1 Testing: Noninvasive Studies

Clinical Diagnosis: Seizure Description

Probably the most important part of the noninvasive studies is the seizure description or semiology. This should lead to a preliminary hypothesis that can be tested and lead to the most probable diagnosis. Special attention should be paid to what both the family and patient report.

Specific auras have been described in different types of seizures. For example, the manifestations of fear and ascending epigastric sensation suggest a mesial temporal epilepsy; simple visual auras suggest an occipital focus; levitation, somatosensory areas; and sympathetic symptoms such as perspiration, difficulty in swallowing, and salivation to an insular focus.¹³

The sequence of symptoms that the patient has is also important; epigastric and psychic auras followed by behavioral arrest and automatisms (ocular, oral, hands, ambulatory) indicate a mesial temporal onset of seizures.¹⁴ A different sequence or visual, auditory, and somatic auras at the beginning point to extratemporal foci with propagation to the mesial temporal lobe.¹⁵ Dystonic positions of the hands as an initial symptom can implicate the contralateral frontal area. If these findings occur late in the seizure event, after behavioral arrest and automatisms, they lose their localizing precision. If a tonic clonic seizure occurs after auras or the symptoms previously mentioned, it is probably the result of propagation of a partial seizure, but if it is not preceded at all by any symptom, then the consideration of a primary generalized seizure should be entertained.

Postictal symptoms have to be taken into consideration; for example, if there is amnesia and sleeplessness for several hours, that suggests seizures of mesial temporal lobe onset. If, on the contrary, there is an immediate recovery, a frontal origin is more probable. Immediate post-ictal dysphasia can also indicate that the seizures may have come from the dominant hemisphere.

Interictal behavioral changes must also be explored since behavioral abnormalities such as perseveration and aggressiveness (frontal lobe symptoms), depression, anxiety, and memory problems (mesial temporal lobe symptoms), can orient toward a specific area. Together with the neuropsychology team, these findings can help design an appropriate set of tests for an individual patient.

Surface Electroencephalogram

The electroencephalogram (EEG) has been considered the most important test for diagnosis since its initial description. Indeed, when abnormal, the EEG provides valuable information:

Interictal Data

Focal discharges consisting of spikes and sharp waves, usually localized to the epileptic area, are suggestive of the focus location. If a focal discharge occurs at or near one of the electrodes common to two channels, one can observe a phase reversal, pointing to the focus localization. If a discharge occurs midway between two electrodes, each will be equipotential and the channel connected between them will not show an abnormality.

Some patients have bilateral secondary synchrony. This occurs when seizure activity starts in a cortical or mesial focus and spreads by the way of the thalamic reticular formation. A careful analysis will show that activity starts in a specific area and then spreads to other areas.

Ictal Data

Ictal recordings usually show acute and rhythmic theta activity, which starts in the affected region and afterwards propagates to other areas. The postictal activity can be slow and unilateral, indicating the affected hemisphere.¹⁶

A normal EEG does not mean that a patient does not have epilepsy; it is also true that the same patient might have different data in several EEG studies recorded on different days. The same patient can have a normal EEG one day, abnormal right-sided data another day, and contralateral data in another recording. This is part of the challenge in finding concordance.

Continuous Video/EEG Monitoring

This technique consists of recording simultaneously EEG as well as the behavior of the patient with a video camera for a period that can last up to several days.¹⁷ This method allows for confirmation of the clinical type of seizure and can correlate it with the onset of abnormal EEG activity, particularly ictal activity. It can also allow for the detection of cases with pseudoepileptic seizures.

Both EEG and video/EEG recordings have their limitations. In spite of the length of the registration period, sometimes seizures do not occur, or multiple movement or muscle artifacts do not permit proper visualization of EEG abnormalities. The spatial resolution is often not specific enough to permit the precise localization of the epileptic zone (EZ) and the clinical seizure type and EEG recording do not always coincide. In many of these cases, intracranial recording becomes necessary. Intracranial recordings can be used to lateralize (identify the side of seizure onset) and/or localize (identify the specific area within a hemisphere) the seizure focus.

Neuropsychological Assessment in Surgical Candidates

The overall objective of neuropsychological assessment in candidates for epilepsy surgery is to seek evidence of cognitive deficits and their relationship with other studies (clinical and electrophysiological).¹⁸^–²⁰

Neuropsychological testing answers specific questions that may impact the surgical decision. What is the patient’s overall neuropsychological status? Which is the hemispheric dominance for language? If cognitive deficits exist, are they lateralized? Is there concordance between neuropsychological findings and structural and electrophysiologic results? Is the patient at risk of a neuropsychological deficit from surgery, and what is the predicted degree of recovery?

There are specific tests for each of the above questions. We will focus on the second and third questions, with particular attention to adult patients with temporal lobe epilepsy.

Hemispheric Language Dominance

Determining the lateralization of language can have a significant impact in surgical planning. It has been established that 95% to 99% of right-handed subjects and 15% to 19% of left-handed have a left cerebral hemispheric dominance for language.^21,²² Due to the small percentage of subjects who may have a right hemisphere dominance, is very important to clarify each case.^23,²⁴ For example, a right-handed patient who presents with difficulties of language might be mistakenly classified as having a left hemisphere dysfunction when in fact he could have dysphasia of the dominant hemisphere, which is not necessarily the left.

For decades, the Wada test has been the gold standard for determining hemispheric language dominance (HLD).^21,²⁵^–²⁷ Recently, a growing number of publications report that other techniques such as functional magnetic resonance imaging (fMRI) may have the same degree of reliability,²⁷^–²⁹ without the disadvantages of cost, risk, and complexity of the technique.³⁰ New noninvasive techniques are evolving to substitute Wada test, however, they still present limitations.^21,^28,^29,³¹

To determine the HLD using fMRI, most authors use the task of evoking words silently, based on a semantic category (e.g., names of animals) to activate temporal regions,³² or a phonologic category (words beginning with the same letter) for activating frontal regions of the dominant hemisphere³³ (Fig. 108-1).

FIGURE 108-1 fMRI study showing left cerebral hemispheric dominance for language in a healthy subject using a phonologic category (see text).

The dichotic listening technique (DLT) has also been proved useful in determining the HLD.^34,³⁵ It consists of simultaneous presentation of two words (one in each ear) with the same characteristics in terms of sound, number of syllables and of common use, with the intention to present some competition for processing stimuli between the two hemispheres. With this methodology it has been shown that the majority of right and left-handed subjects have right ear advantages as a reflection of a left HLD.³⁴ Fig. 108-2 shows the variability in the index of HLD according to the dichotic listening test in right-handed healthy subjects. The lateralization index for language (LI) is a predictor of postoperative verbal memory deficit as shown by some studies.³⁶

FIGURE 108-2 Variability in language laterality in right-handed healthy subjects using the dichotic-listening test. The LI is obtained as follows: (L – R)/(L + R) = LI, where L = number of words heard with the left ear, and R = number of words heard with the right ear. A positive LI corresponded to a left-predominant lateralization and a negative LI corresponds to a right-predominant lateralization. B, bilateral representation (+0.25 to –0.25); LI, laterality index; SLL, strong left lateralization (+0.50 to +1); SRL, strong right lateralization (–0.50 to –1); WLL, weak left lateralization (+0.25 to +0.50); WRL, weak right lateralization (–0.25 to –0.50). (Redrawn from Lehéricy S, Biondi A, Sourour N, et al. Arteriovenous brain malformations: is functional MR imaging reliable for studying language reorganization in patients? Initial observations. Radiology. 2002;223(3):672-682.)

Lateralization of Memory Deficits

Lateralization of memory deficits (LMD) is another objective of preoperative neuropsychological assessment. Table 108-1 ³⁷ presents some of the available memory tests. At our clinic, we use the Battery of Learning and Memory and the Visuospatial Learning and Memory tests. When evidence is found for a memory deficit in the same hemisphere where the epileptic zone has been localized through clinical, EEG, and imaging data, the decision for surgery is strengthened; otherwise, the complete testing described in Table 108-1 is applied. Evaluation of memory is performed by well-established tests that vary slightly from one center to another.³⁷ In approximately 50% of patients it is not possible to determine the lateralization of memory deficit through neuropsychological studies alone;¹⁸ however, that percentage decreases when the information provided by neuropsychological testing is concordant with other studies. A major problem in LMD is that there is a high error rate due to language dominance, as discussed above; that is, verbal memory deficit does not always indicate dysfunction of the left cerebral hemisphere.

Table 108-1 Most Frequently Used Tests in Epilepsy

General BatteriesWechsler Intelligence ScaleHasltead-Reitan BatteryHemispheric lateralizationWada TestDichotic ListeningFunctional Magnetic Resonance ImagingAttentionTrail Making TestCancelation TestLanguageBoston Diagnostic Aphasia ExaminationBoston Naming TestToken TestVisuospatial and PerceptualHooper Visual Organization TestConstructional ApraxiaBenton Judgement of Line OrientationMotor and Reaction TimeFinger OscillationHand DynamometerGrooved Pegboard

Problem Solving, FlexibilityWisconsin Card Sorting TestWord FluencyStroop TestBattery of Learning and MemoryWechsler Memory ScaleVerbal Learning and MemoryStory RecallPaired Word LearningRey Auditory Verbal Learning TestCalifornia Verbal Learning TestVisuospatial Learning and MemorySimple Designs RecallRey Osterrieth Complex FigureBenton Visual Retention TestOthersBeck Depression InventoryQuality of Life in Epilepsia (QOLIE-31)International Neuropsyquiatric Interview (MINI)

Imaging Studies

The best imaging method to study intractable epilepsy is magnetic resonance imaging (MRI) because of its excellent spatial resolution employing basic sequences such as T1WI, T2WI, and FLAIR. It has high sensitivity for characterizing signal intensity of normal and pathologic brain tissue and very good specificity for neoplastic, vascular, atrophic, dysplastic, infectious, and degenerative etiologies. MRI allows the differentiation of edema, demyelinating diseases, heterotopic gray matter (Fig. 108-3), and space-occupying lesions involving anatomic structures causing epilepsy.

FIGURE 108-3 A and B, Axial and coronal images, showing migration and sulcation disorders, with gray matter nodular heterotopy. Note the abnormal localization of gray matter in deep white matter lateral to right ventricle with ependymal extension (arrows).

In a series of 40 patients with refractory focal epilepsy, Knake et al.³⁸ found that studies with 3T phased-array surface coil (PA-MRI) yielded additional diagnostic information in 48% of the studies (19/40), when compared to routine clinical MRI performed at 1.5T. In the subgroup of patients with previous 1.5T MRIs interpreted as normal, 3T PA-MRI resulted in the detection of a new lesion in 65% of patients (15/23). In the subgroup of 15 patients with known lesions, 3T PA-MRI better defined the lesion in 33% (5/15). On the other hand, Zijlmans et al.³⁹ stated that patients studied with 1.5T show loss of cerebral tissue and mesial temporal sclerosis better than at 3T, while those patients with cerebral dysplasias are better studied with 3T. High-resolution 3T MRI (HR 3T MRI) and surface coils applied over the suspected epileptogenic zone are useful to detect lesions in patients suffering refractory epilepsy due to cortical developmental malformations (CDM).⁴⁰

In order to optimize the diagnosis of mesial temporal sclerosis and the severity of hippocampal atrophy, it is important to obtain a precise hippocampal evaluation with axial and coronal images in MRI. Axial images oriented perpendicular to the axis of the clivus, 2-mm thick without any gaps, provide an adequate view of the hippocampal structures. Coronal images are taken perpendicular to axial sections. Hippocampal areas can be evaluated with these regions of interest (ROI) measurement to define possible differences in volumetric areas indicating hippocampal atrophy (Fig. 108-4A). Flair sequences offer an objective method to evaluate hippocampal area signal intensity in order to define hippocampal atrophy and mesial temporal sclerosis (Fig. 108-4B) that may be associated to ipsilateral mammillary body and fornix atrophy (Fig. 108-5). Successful surgery is possible even with normal MRI but it requires compelling clinical and electrophysiologic evidence of seizure onset.

FIGURE 108-4 A, T2 coronal image with region of interest measurement of the hippocampus. There is atrophy and hyperintensity on the left due to hippocampal sclerosis (arrow). B, Flair sequence of coronal image showing left hippocampal hyperintensity (arrow), corresponding to sclerosis and atrophy.

FIGURE 108-5 T1 coronal images. Left hippocampal sclerosis is associated with widening of temporal horn and ipsilateral atrophy of fornix (A) and mammillary body (B) (arrows).

Magnetic Resonance Spectroscopy

While MRI is primarily employed to obtain structural images of the brain, magnetic resonance spectroscopy (MRS) can assess regional cell loss through determination of the concentrations of intermediate metabolites, including glutamate and glutamine.⁴¹ This noninvasive technique can also measure other metabolites of cellular activity such as N-acetyl-aspartate (NAA), creatine (Cr), and choline (Ch) that give indirect information regarding cellularity of the ROI. MRS provides also a broad range of useful functional information such as cerebral concentrations of GABA and glutamate, usually associated with an increase in pH and inorganic phosphate and reduction of phosphate monoesters. Several studies in patients with epilepsy have documented neuronal loss and alterations in energy and lipid metabolism, acid–base homeostasis, and amino acid neurotransmitter metabolism.⁴² Proton MRS imaging studies consistently demonstrate decreased NAA in the epileptogenic temporal lobe (Fig. 108-6).

FIGURE 108-6 Hippocampal univoxel spectroscopy. There is bilateral reduction of N-acetyl-aspartate compatible with hippocampal sclerosis.

Functional magnetic resonance imaging (fMRI) is a noninvasive functional brain mapping technique assessed on blood oxygen level–dependent (BOLD) signal with echo planar images (EPI) obtained during T1-weighted imaging MRI studies. It offers a map of physiologic and metabolic functions of cerebral activity during ictal and interictal discharges on images with a spatial resolution of a few millimeters and a temporal resolution of a few seconds.^43,⁴⁴ One of the first clinical applications of fMR was presurgical evaluation of cerebral function in patients with epilepsy and neoplastic lesions nearby eloquent areas. This technique detects the localization of the functional areas of primary sensorimotor cortex or language zones prior to surgery. fMRI offers noninvasive preoperative brain mapping with high sensitivity to detect cerebral lesions and defining the border between lesion and normal functional cortex. It may also predict possible deficits in motor, sensory, or language functions due to expansion of the lesion or surgical procedures. Therefore, it offers information to help therapeutic decisions relative to risk–benefit ratio of the treatment. In patients with refractory epilepsy fMR has been used to evaluate the resection feasibility, planning the surgical procedure and better patient selection for invasive mapping⁴⁵ (Fig. 108-7). Functional MRI holds great promise as a powerful tool in memory evaluation.³¹ fMRI aids in the localization of language and motor function of candidates for epilepsy surgery, and has up to a 90% concordance with WADA test (Fig. 108-1).⁴⁶

FIGURE 108-7 Functional magnetic resonance imaging (fMRI) and electrophysiologic brain mapping (BM) in a case with partial removal of low-grade astrocytoma, used to plan total surgical resection. A, Tridimensional MRI reconstruction showing the proximity of previous resection cavity (upper arrow) and residual tumor (lower arrow) to the central fissure. B, Superimposition of MRI, fMRI, and subdural 36–contact grid user for BM. Black zone indicates the place of previous resection, grey zone the tumor mass to be removed, and red zone the area of activation in fMRI using a contralateral hand motor paradigm. Circles correspond to contacts of subdural grids used for BM. Empty circles indicate contacts used silent to electrical high-frequency stimulation (130 Hz), yellow dots contacts from where sensory responses were elicited, and black dots contacts from where somatic-evoked responses were recorded by stimulation of contralateral median nerve. White circles indicate the electroencephalogram-epileptic zone. BM, brain mapping; fMRI, functional magnetic resonance imaging.

Diffusion Tensor Imaging

Diffusion tensor imaging (DTI or tractography) allows the detection and examination of the integrity and orientation of discrete white matter fiber bundles not optimally evaluated with conventional MRI. Recently DTI/tractography has been applied to the study of epilepsy and has demonstrated diffusion changes in gray and white matter. DTI gives information on the path of axonic fibers in the white matter as a support to surgical planning.⁴⁷ It is particularly useful in determining the location of visual pathways in cases proposed for mesial temporal lobe resection and in assessing motor pathways for lesions in the perirolandic area (Fig. 108-8). Intractable TLE is marked by widespread involvement of fiber tracts and asymmetries of white matter fibers, with lower fiber fractional anisotropy (FA) ipsilateral versus contralateral to the seizure focus.

FIGURE 108-8 Tractography of horizontal temporal-frontal and temporal-occipital (Meyer’s) tracts in a patient with left anterior temporal lobectomy associated with quadrantanopsia. On the side of surgery, there is interruption of fibers of Meyer’s loop (arrow).

Positron Emission Tomography

The primary use of PET in epilepsy is presurgical localization of the epileptic focus in patients with complex partial seizures. (CPS). It also quantifies cerebral metabolism, blood flow, oxygen extraction, and receptor kinetics in an attempt to understand the pathophysiology of epilepsy.⁴⁷ The positron-emitting radionuclides used most commonly in PET imaging are oxygen-15 (¹⁵O), nitrogen-13 (¹³N), carbon-11 (¹¹C), and fluorine-18 (¹⁸F).⁴⁸ Assessment of glucose metabolism using ¹⁸F-2-deoxyglucose (FDG) and cerebral blood flow using H₂¹⁵O (^O15) are the two most common physiologic processes measured by PET. Most PET images are the result of time-dependent integration of metabolic activity. For FDG studies, this varies from 30 to 45 minutes.⁴⁹ FDG PET has the highest sensitivity and specificity in nonlesional epilepsy by acting as a combined measure of metabolism and anatomy.⁵⁰

Single-Photon Emission Computed Tomography

Ictal SPECT has been extensively studied in patients with partial epilepsy driven primarily by the need for preoperative localization, the volume of patients affected, and the type of seizure activity. The major limitation to ictal SPECT is the need to inject the tracer within a minute or so of seizure onset. In the temporal lobe there is a pattern of markedly increased perfusion accompanying the ictus, involving the mesial and often to a greater extent the anterolateral neocortical structures.⁵¹ Perfusion follows changes in metabolism during seizures. Imaging interictal and ictal brain-blood flow with SPECT is a feasible alternative to PET, requiring less capital outlay, using inexpensive perfusion tracers, and being readily available.⁵²

Phase II Testing: Invasive Studies

When noninvasive procedures fail to determine the anatomic origin of partial epilepsies, either because results are equivocal or there is no concordance among them, invasive studies become necessary. This more often occurs in:

1. Patients without structural lesions in imaging studies.

2. Patients with bilateral homologous interictal discharges and/or alternative onset of ictal activity in both hemispheres in video EEG studies.

3. Discordance among clinical seizure type, EEG, and image studies on the localization of the epileptic zone (EZ).

4. EZ located nearby eloquent areas which may compromise surgical resection and/or pose a risk of a neurologic deficit.

Sphenoidal Electrodes

This minimally invasive procedure, performed under local anesthesia and guided by fluoroscopy, intends to bring a percutaneously-inserted electrode tip close to the bone window represented by the foramen oval at the skull base.

With the patient positioned in dorsal decubitus, head hyperextended, and the x-ray beam angulated to take a skull base film, a 21-gauge, spinal tap needle is inserted 2 cm lateral to the lip commissure. Guided by fluoroscopy, the needle is directed to the entrance of the foramen oval. A fine stainless steel electrode, insulated except at its tip, is passed through the needle just to the level of the skull base. The needle is removed and the electrode left and taped in place to the cheek (we use Micropore tape, 3M, St. Paul, MN).

Sphenoidal electrodes help to define the side of the mesiotemporal epileptic foci (MTE) in cases of equivocal scalp EEG recordings. They are sometimes used to rule out MTE in cases with pseudo epileptic seizures (Fig. 108-9). The reason for limiting the use of sphenoidal electrodes is that unfortunately EEG recordings can be contaminated with various artifacts and are uncomfortable for the patient.

FIGURE 108-9 Skull base radiogram showing implanted sphenoidal electrodes. The tip of the inserted electrodes is at the entrance of the foramen ovale (FO), which is placed anterior and medial to the foramen rotondum minor (FR). Recordings from the electrodes clearly show unilateral interictal spiking on the right sphenoidal electrode leads (R SPH), when compared to scalp monopolar leads (A1–A2) and the left sphenoidal (L SPH).

Intracranial Electrodes

Intracranial electrodes are used to explore cortical and subcortical structures by recording spontaneous and evoked EEG activity and applying electric current to obtain clinical and EEG responses. Intracranial electrodes may be of two types (Fig. 108-10):

FIGURE 108-10 Different types of intracranial electrodes. 1, Sphenoidal electrodes. 2, Grids and strips consisting of silicon plates containing round plate contacts, to be inserted subdurally. 3, Parasagittal intracerebral electrodes to explore longitudinal areas such as the amygdala-hippocampus complex. 4, Orthogonal placed electrodes to explore different cortical and subcortical structures in suspected multifocal epilepsy involving more than one cerebral lobe.

Plate Electrodes

They come in the form of flexible grids containing from 20 to 64 rounded contacts or strips with 4 to 16 contacts. Plate electrodes are placed subdurally and are used to explore cortical areas in the convexity, base of the brain, and interhemispheric cortex. Their placement requires the exposure of the cortical area to be explored, therefore, the size and location of the initial craniotomy are of paramount importance. When the epileptogenic area has been localized over the convexity by noninvasive studies, the placement of grids and strips to define its extension and proximity to eloquent areas requires a craniotomy large enough to allow for direct visualization of all or as much of the grid’s contacts as possible. The worst scenario is that the epileptic focus is beyond the craniotomy’s edge, forcing the surgeon to extend the incision or to make a new one perpendicular to the original, thus creating acute edges and compromising skin and/or bone flap vascular supply.

The craniotomy size and location and the size of the grid or strips depend mostly on the results of presurgical evaluation studies, including ictal semiology, surface EEG, and imaging studies. Obviously, basal and mesial areas of the brain are difficult to see directly regardless of craniotomy size. When the epileptic area is placed in cortical basal areas, subdural grids are slipped under the brain through a craniotomy of a size that allows an entire lobectomy. This also applies for the placement of interhemispheric grids, which are often bilateral.

We do not recommend burr-holes to introduce strips of electrodes over the cerebral surface since they provide a poor spatial resolution for epileptic foci location and brain mapping.⁵³ For both strips and grids, subdural bridging veins or adhesions could constitute a problem for accurate placement.⁵⁴

Deep Brain Electrodes

Deep brain electrodes are fine tubular (0.8–1.3 mm in diameter) devices with multiple contacts along their trajectories. They are inserted in different cortical and subcortical anatomic structures using stereotactic techniques, guided by different imaging modalities (x-ray ventriculography, angiography, CT, and MRI), and coupled with anatomic atlases by a fusion imaging software. Imaging fusion allows elaborating virtual trajectories to reach the targets precisely and traversing non vascular territories ⁵⁵ (Fig. 108-11). Deep brain electrodes may be inserted through parasagittal frontal, parietal, and occipital approach, or perpendicular to the sagittal plane (orthogonal approach).

FIGURE 108-11 Imaging fusion of perioperative computed tomography (lower half) showing the planned trajectory to insert an amygdala-hippocampal electrode (interrupted white lines). This image is superimposed to a postoperative magnetic resonance image (upper half), showing the electrodes in place (black dots corresponding to electrodes contacts), in both the amygdala (anterior arrow) and the hippocampus (posterior arrow).

In the parasagittal approach, electrodes usually traverse greater distances to get to the subcortical targets; however, their trajectories may be easily planned to avoid vascular areas. This approach is used to explore longitudinal anatomic areas, such as the amygdala-hippocampal complex when bilateral independent epileptic foci in this region are suspected ⁵⁶ (Fig. 108-12). Frontal and parietal parasagittal approaches are used to explore the insular cortex, avoiding the risk of orthogonal placement traversing the densely vascular Sylvian fissure.^13,⁵⁷

FIGURE 108-12 Bilateral parasagittal electrodes inserted through occipital approach to explore the amygdalae and hippocampi in a case suspected to have bilateral hippocampal independent foci. A, Coronal section showing electrodes within the hippocampus (arrows). B and C, Sagittal and axial sections showing the position of the two most anterior contacts in the amygdala (anterior arrows) and the last four contacts within the hippocampus (posterior arrow).

An orthogonal approach allows a better definition of the EZ and areas involved in the propagation of the epileptic seizure, either unilateral or bilateral. It is the most adequate method to be used in cases where multilobar participation in seizure genesis is suspected, such as in frontal-temporal, frontal-parietal, and temporal-occipital foci, as an almost unrestricted number of electrodes may be inserted with low morbidity.⁵⁸

Through intracranial electrodes, we record EEG from different structures of the brain that may be involved in the genesis and propagation of spontaneous seizures.⁵⁸ The advantage of recording with intracranial electrodes is that they are not subject to physiologic artifacts such as movement and sweating, they are much closer to the seizure origin, and their spatial resolution is high and can show ictal onsets with specificity.⁴ When EEG recordings with intracranial electrodes are performed with simultaneous videotelemetry, we can determine the relationship between behavioral changes and EEG data with greater accuracy (Fig. 108-13 and Video 108-1).

FIGURE 108-15 Video/EEG continuous monitoring of patient KG 63. This patient has a left temporal basal grid with 20 contacts. To the right is the patient who is peaceful, with no visible seizure clinical symptoms. He holds a button in his right hand that he presses to signal that he is having an aura, but cannot talk. To the left is the EEG recording; the first four channels are recording from the left surface frontotemporal region and the rest correspond to the 20 electrode contacts from the grid; all contacts are referred to the left ear. White arrow indicates where the ictal EEG activity starts. Low-voltage fast activity appears in contact 1 (corresponding to the left anterior parahippocampus). This activity begins 6 seconds prior to propagation to other contacts (not shown). Note that no activity is recorded in the first four channels corresponding to the surface EEG.

Interictal Data

Figure 108-13 shows the interictal spikes that are characteristic of the epileptic tissue. Unlike surface EEG, interictal elements detected with intracranial electrodes are very conspicuous, with high amplitudes. The discharges are acute and fast. It is frequent to find interictal spikes in several contacts and as such, it is difficult to base our focus localization without ictal activity, although sometimes they have high amplitude and show phase reversal which suggests the precise focus location as seen in the figure. Sleep studies are also useful since during REM sleep, the interictal spikes are more localized to the seizure onset.

FIGURE 108-13 Interictal data in a patient with a right basotemporal grid. The first three channels correspond to surface electroencephalogram and the rest to the temporal grid. Activity is restricted to two channels with phase reversal in contact six corresponding to anterior parahippocampus. Note that surface recording does not show epileptic activity.

Ictal activity is the gold standard to define focus location. In the hippocampus, onset of seizures is anticipated by either desynchronized fast activity occurring in the epileptic focus ⁵⁹^–⁶² or a high-voltage slow pattern.⁶³ When ictal onset is found in one or two contacts, it is considered a focal onset; if on the other hand, ictal activity onset is observed in more than three contacts simultaneously, it is not a focal but a regional onset.⁶⁴ If, as shown in Fig. 108-13 and Video 108-1, the ictal EEG onset is focal and precedes for several seconds the behavioral changes, the postsurgical results regarding seizure reduction are better.¹²

Epilepsy centers are now including the identification of high frequency oscillations (HFOs), known as ripples (80–250 Hz) and fast ripples (250–500 Hz) as part of the process of identification of the epileptic focus. Total rates of HFOs are significantly higher in the seizure-onset zone than outside, their rate either increases or decreases before the onset of seizures and they extend beyond the epileptic focus, although their pattern has not been described. There appears to be no association between the extent of the area from where HFO are recorded and seizure count but an association has been found between the rate of HFO and seizure count.⁶⁵^–⁶⁷

On the other hand, electric current applied to different contacts of intracranial electrodes may serve to identify the EZ as the place with lowest threshold to induce an EEG after discharge, possibly accompanied by a clinical seizure similar to the spontaneous seizures the patient presents (Fig. 108-15).⁶⁸

FIGURE 108-14 After discharge elicited by bilateral electrical stimulation of adjacent contacts, where the epileptic focus had been identified at the onset of spontaneous seizure, recorded from an electrode placed along the amygdala and hippocampus. Electrical stimulation was performed with a pulse of 130 Hz, 1.0 msec pulse width, 630 μA of amplitude, in a train of 10 seconds. The after-discharge initiated simultaneously in the amygdala and hippocampus, near the place of stimulation, and propagated to other areas as seen by scalp EEG. Propagation of the EEG seizure was accompanied by a clinical seizure similar to the ones the patient presented spontaneously. EEG, electroencephalogram.

Intracranial Electrodes and Brain Mapping

Electric current applied through intracranial electrodes is used to identify the location of eloquent areas. BM is carried out with the patient awake and enables the mapping of motor and speech areas.⁶⁹ The spatial precision is excellent particularly with low intensities.

Two types of pulses are used to map eloquent areas: low frequency (1–2 Hz), high amplitude (1.0–15.0 mA), and long-pulse duration (1.0–3.0 msec) electric current is used to induce focal muscular contractions in contralateral extremities when applied to motor cortex. High-frequency (50–130 Hz), low-amplitude (0.5–10.0 mA), and short pulse-width (0.21–1.0 msec) electric current interferes with spontaneous ongoing activity, such as speech or voluntary movements, or elicit sensory changes in the form of paresthesias.⁷⁰

Intraoperative versus Extraoperative Electrocorticography and Brain Mapping

Electroencephalogram recording and brain mapping (BM) may be performed just before resection of the EZ or the craniotomy for the electrode implantation can be performed in a first operation and resection of EZ as a separate procedure.⁷¹

Although intraoperative electrocorticography (ECoG) and BM have the obvious advantage of reducing the study and treatment to a single procedure, thus given the impression of being safer, in our experience the inconveniences outweigh the advantages. At our institution we do not use intraoperative ECoG but rather an extraoperative long-lasting recording for a number of reasons:

1. Time available to record electrical cortical activity is limited during surgical procedures, resulting on a limited chance to record a seizure. More often one records only interictal epileptiform activity.⁷²

2. The neurophysiologist must read and interpret the EEG recordings and BM as they are obtained, thus mandating the need for both the equipment and neurophysiologist to be either in the operating room or in an area where close communication with the surgeon is possible.⁷² If cognitive functions or language has to be explored, the presence of the neuropsychology team is also required. This often results in crowded operating areas where the sterility of the procedure is compromised and accidents are likely to occur (e.g., tripping over one of the many cables).

3. Spontaneous seizures seldom occur during intraoperative ECoG, so important decisions have to be taken based on interictal recordings or on seizures induced by a convulsing drug. Seizures induced with pentilenetetrazol or other methods do not necessarily originate at the epileptic focus. Besides, spontaneous seizures that may occur during the intraoperative electrocorticography are not necessarily the ones that the patient presents most frequently.

4. Intraoperative ECoG recording is usually done with an awake patient because of concerns of interference of anesthetic agents with the recordings. It is noteworthy that sufentanil, fentanyl, alfentanil, propofol, and methohexital have been reported to produce epileptiform changes on ECoG, whereas propofol, halothane, barbiturates, and benzodiazepines may suppress the epileptic activity.^72,⁷³

5. Awake craniotomies are particularly complicated in children and in developmentally disabled, anxious, or even tired patients. The surgical experience can generate unnecessary stress and poor cooperation that is only augmented by the cortical mapping phase of the procedure.

Surgical Technique for Subdural Electrode Placement

Standard surgical techniques are used; general anesthesia is undertaken. Bony surfaces should be padded as necessary; microsurgical techniques are seldom necessary so head fixation is not mandatory. The skin flap is reflected and a bone flap created, making sure that thorough hemostasis is obtained; placement of an epidural hemostatic agent below the craniotomy’s edges and dural tack-up sutures should not be missed and once the dura has been opened the surgeon must look for and coagulate bleeding points on the dural edge. All of these precautions tend to prevent a subdural hematoma that could interpose itself between the arachnoid surface and the electrode grids resulting in movement or faulty recording from the latter.

Electrode grids and strips should be soaked in saline solution to allow for an easier sliding over the arachnoid surface, particularly beyond the craniotomy’s edges. It’s of utter importance that the surgeon identifies perfectly each cable of the grids used to ensure that proper connections with the EEG head mount and stimulator are established. Our practice is to write down the color codes of each cable and the cortical area they cover and use specific suture materials to tie together the cables of the same grid. Grids are fixed in position with a 4-0 nylon suture from their edges to the dura. A drawing or sketch of the contacts of the grids and strips in relation to identified cortical landmarks helps the neurophysiologist in locating the epileptogenic and eloquent areas.

The dura should be as close to watertight as the cables allow. We do not use stab wounds to externalize the cables to prevent contamination of the tunnel when the distal end of the cables is pulled for removal. Rather, we externalize them through the surgical wound. Galea and subdermal tissue are closed with an absorbable suture and the skin is closed with a nonabsorbable one.

After surgery the patient is taken to his or her room and a postoperative MRI confirms electrode location. Antibiotics are continued throughout the entire process of focus identification and cortical mapping; antiepileptic drugs are usually discontinued the night of the surgery so that recording can begin the next day. Patients with a history of status epilepticus, abundant seizures, or severely altered surface EEG are not taken off medications until we have evaluated the first days of recordings and the need to gradually discontinue the medications. Head dressings are changed daily, giving us the chance to inspect the wound. When a cerebrospinal fluid (CSF) fistula occurs, additional stitches, head elevation, bed rest, and sometimes acetazolamide are used.

Close neurologic observation is mandatory since patients with subdural grids may develop subdural hematomas (see complications below). Should a patient present three or more consecutive generalized seizures in close proximity, we administer 5 to 10 mg of diazepam to prevent status epilepticus.

Surgical Technique for Intracerebral Electrodes

Intracerebral electrodes are implanted using stereotactic techniques. After the frame has been placed, a contrast-enhanced CT or MRI is obtained. We prefer to obtain both and use image fusion to analyze the stereotactic position of the targeted structure. This gives the advantage of indirect calculation using the CT image, which is not subject to distortion and the anatomic identification of the target using the more detailed MRI. Moreover, after CT and MRI have been fused, the target is analyzed on anatomic atlases provided with the fusion software.

For implanting the electrodes to explore the amygdala-hippocampal complex using an occipital approach, two burr holes are placed guided by the stereotactic trajectory. Electrodes are implanted and held in place with a plastic ring that fits the size of the burr hole and a cap. They are externalized through the wound to prevent contamination when explanting them.

Orthogonally implanted electrodes require a much more sophisticated technique. CT and MRI images should be fused with an angiography to prevent injury of vascular structures. Each electrode has four to eight contacts and at least three of them are used to explore the hippocampus and amygdala. They offer the advantage of recording neocortical and paleocortical structures simultaneously, thus allowing the detection of extramesial foci. Other areas of the brain, anatomically related to the suspected EZ, can be covered. This is the case of orbitofrontal and motor areas. Insular electrodes can be implanted but, as mentioned previously, due to the high risk of injuring the Sylvian vessels, orthogonal electrodes have been substituted by parasagittal, frontally, or parietal inserted electrodes. Each orthogonal electrode requires a stereotactically planned burr hole and a separate stab wound. Moreover, when extraoperative recording is to be undertaken, implantation of a screw in the cranial vault is used to hold the electrodes in place. Hollow macroelectrodes allow the implantation of microelectrodes in the same trajectory for unit recordings, and are used mainly for research purposes. In some cases restricted EZ or lesions causing epileptogenesis may be destroyed by radiofrequency lesions through the same orthogonal electrode used for recording.⁷⁴

Focal Epilepsy Detected Near Eloquent Areas

Cortical stimulation is usually indicated when the planned resection will be in or adjacent to eloquent cortex. ECoG and BM allow the identification of epileptic foci in or near eloquent cortical areas and permit tailoring the resection with satisfactory safety and seizure relief, this has been proved for pure insular lesions and those extending to adjacent lobes.⁷⁵ In sensory-motor cortices, an ECoG that demonstrates few postresectional spikes and no distant spikes predicts a better seizure control.

Complications

Infection rate has been reported as 3.9% and epidural hematomas and brain edema rates in 2% and 8%, respectively.⁷⁶ It has been said that almost all patients subject to subdural grid placement for extraoperative recordings develop a subdural collection,⁷⁷ but it has also been reported that only 7.8% of them require surgical drainage.⁷⁶ Most require only conservative management, since volume, midline shift, and maximal thickness do not predict the clinical course of this patients, so sound clinical judgment must be exercised to guide these patient’s care and need for evacuation.⁷⁷ The artifact caused by electrodes on CT makes evaluation of electrode placement and complications difficult and MRI is the preferred method to evaluate them.⁷⁸

In a series of 67 pediatric patients, a CSF leak was seen in 21 patients, 10 had positive subdural cultures but only 1 of these had a positive lumbar CSF culture and none developed clinical meningitis, and 1 patient developed transient visual field loss after placement of grids over the occipital lobes.⁷⁹ Another series of 112 children (122 procedures) revealed that placement of additional electrodes was necessary in 5.7% of patients; wound infection occurred in 2.4% of these patients; cerebrospinal fluid leak in 1.6%; and subdural hematoma, symptomatic pneumocephalus, bone flap osteomyelitis, or strip electrode fracture occurred in 0.8% each. There were four cases of transient neurologic deficit (3.3%) and no permanent deficit or death.⁸⁰