3.1 Introduction

Epilepsy is a common neurological disorder with a prevalence of 0.5 to 1%. ¹ , ² , ³ Furthermore, 20% of patients have seizures despite treatment with maximally tolerated medication. ⁴ , ⁵ , ⁶ , ⁷ Estimates are that approximately 3 to 12% of epilepsy patients and 30 to 50% of pharmacoresistant patients may be candidates for surgical treatment. ⁸ , ⁹ Traditionally, these surgeries have involved brain resections targeting the presumed region of seizure onset, often referred to as the epileptogenic zone (EZ). A patient with medically refractory epilepsy undergoes an extensive evaluation (see below) to determine if he or she is a candidate for such a surgical approach. Some patients are recommended for a resective surgery on the basis of this evaluation. Resective surgery might involve intraoperative testing, such as stimulation testing to map eloquent cortex and/or electrocorticography to identify suspected epileptogenic cortex, to tailor the resection. For many patients, the preoperative evaluation suggests that further testing could lend support for a resective surgery or other surgical intervention (e.g., implantation of a neuromodulation device, such as Responsive Neurostimulation [RNS]). This further testing involves evaluation with surgically implanted intracranial electrodes.

Surgical treatment of epilepsy requires the precise location of the seizure foci to maximize the potential for surgical remission and minimize disruption of normal brain function. The most common cause for failure in the surgical treatment of intractable epilepsy is incomplete surgical resection of the EZ. ¹⁰

Some criteria used by the surgeon to consider the use of invasive techniques instead of relying on noninvasive technologies are lack of lateralization or localization of the EZ, discordant information on noninvasive studies, the presence of the EZ near eloquent cortex, and the need to define the relationship of the EZ to a lesion present on neuroimaging. ¹¹

Intracranial electrodes permit extraoperative continuous intracranial electroencephalographic (iEEG) investigation. This allows abundant collection of interictal data, recording of seizure onsets, and seizure spread patterns, and may (depending on electrode type and location) permit stimulation mapping of cortical function. Although cortical mapping can be performed in the operating room, extraoperative mapping using implanted electrodes offers several advantages. First, more complex cognitive functions can be mapped in implanted patients, in comparison with the limited tasks that can be mapped in the operating room. Second, in contrast to intraoperative mapping, which is generally limited to 1 to 2 hours, implanted electrodes usually remain in place several days, permitting significantly more time for testing. Third, iEEG offers a means of brain mapping in patients who are unable to cooperate with an awake craniotomy. The acquired data may be used to decide upon a resective surgery or upon sites for implantation of electrodes to be used with an RNS device. Electrode types include epidural, subdural, and depth electrodes. Depth electrodes are addressed elsewhere in this book. Unlike scalp EEG in which a significant amount of tissue is interposed between the recorded cells and the electrode, electrodes placed by invasive methods localize the electrode on or near the cells that are being recorded, resulting in less attenuation of the signal and better signal-to-noise ratio. The amplitude of the potentials recorded from the cortex using invasive techniques is typically 2 to 58 times greater than scalp recordings. ¹²

This chapter will address the application of epidural and subdural electrodes, including review of their development, selection of patients for implantation, strategy for implantation, operative technique considerations, and patient outcome.

3.2 History

The implantation of intracranial/invasive electrodes was first performed in 1939 by Wilder Penfield and Herbert Jasper at the Montreal Neurological Institute using epidural electrodes. ¹³ , ¹⁴ Scalp EEG performed on the patient demonstrated bilateral temporal epileptiform discharges. Bilateral epidural electrodes were implanted and recording identified epileptiform activity on the patient’s left side. An awake left temporal craniotomy revealed a posterior temporal meningocerebral scar that was then resected. In 1949, Hayne and Meyers reported the first use of stereotactically implanted depth electrodes for epilepsy monitoring. ¹⁵ Around the same time, Talairach and Bancaud began using stereotactically implanted depth electrodes (stereoelectroencephalography [SEEG]) and became the first center to consistently utilize intracranial ictal recordings to guide resections of epileptic foci. ¹⁶ Although subdural electrodes were implanted in some patients as early as the 1950s, ¹⁷ it was not until the 1980s that implanted subdural strips and grid arrays of electrodes became popularized. ¹⁸ In this chapter, we will focus on the implantation of cortical surface electrodes (peg electrodes, subdural strips, and subdural grids) in the evaluation and treatment of patients with medically refractory epilepsy.

3.3 Patient Selection

Epilepsy patient evaluation typically includes video/electroencephalogram (EEG) monitoring, magnetic resonance imaging (MRI) tailored to look for structural abnormalities known to be associated with focal seizure onset, interictal positron emission tomography (PET), ictal single-photon emission computed tomography (SPECT), magnetoencephalography (MEG), a detailed neuropsychological evaluation, and intracarotid amobarbital procedure (IAP) or Wada test. Many centers also utilize functional MRI. Typically, patients are considered potential candidates for epileptogenic focus resection if their seizure semiology and scalp EEG findings are consistent with origin from a single brain region. Among centers with extensive experience with the treatment of these patients, the choice of surgical approach varies. In every center, a number of patients are recommended for resective surgery on the basis of a concordant noninvasive evaluation. In general, in addition to seizure semiology and scalp EEG findings that are consistent with origin from a single brain region, these patients have a solitary structural abnormality on brain MRI that is consistent with being an epileptogenic lesion. Common MRI abnormalities include cavernous malformation, unilateral hippocampal sclerosis, or a suspected low-grade neoplasm or hamartoma. These patients undergo a preoperatively determined resection (e.g., temporal lobe resection) or a resection guided by intraoperative electrocorticography (ECoG). For patients who are thought to have a resectable epileptogenic focus that cannot be adequately identified by the noninvasive evaluation, an invasive electrode investigation is pursued with the goal of identifying a resectable focus, and/or to identify functionally important cortex to be preserved. These patients typically do not have a solitary structural lesion identified on brain MRI. Alternatively, patients may have a solitary structural lesion, but a scalp EEG seizure-onset zone localized to a different brain region. Patients without an identified lesion often have scalp EEG monitoring that has either not clearly lateralized their seizure onsets or, if lateralized, not adequately localized their seizure onsets to a specific lobe or part of a lobe.

3.4 Electrode Types

There are several types of surface electrodes that differ in their location in the cranial space, as well as their configuration. The choice of invasive monitoring depends on patient characteristics including the site of disease, the age of the patient, and results of preoperative testing. Peg electrodes are mushroom-shaped electrodes consisting of a steel or platinum disk with a Silastic cover (▶Fig. 3.1). These electrodes are normally placed via twist drill into the epidural space. The main clinical indication for peg electrodes is lateralization of seizure focus. These electrodes are used in patients in whom scalp EEG is not possible because of anatomical reasons (e.g., bone defects) or in patients in whom scalp EEG failed to lateralize the seizure activity. ¹⁹ , ²⁰ Peg electrodes do not record directly on the cortical surface; therefore, they have limited ability to spatially define the EZ. Also, because of their shape and configuration, they cannot be used to monitor subfrontal or subtemporal regions. However, modified peg electrodes with cylindrical forms have been used successfully in these regions. Because the dura is left intact, these electrodes are believed to have a reduced rate of complications such as hematoma and infection.

Strip electrodes consist of a linear array of platinum or stainless steel contacts embedded into a silicon strip (▶Fig. 3.2a). These constructs normally consist of four to eight contacts spaced 10 mm apart. These electrodes are placed into the subdural space either through a burr hole or under the edge of a bone flap. This type of electrode is used both for intraoperative monitoring during tumor resections and for postoperative monitoring in seizure cases. These electrodes have some advantages over larger grid electrodes. Because of their narrow size, they can be passed considerable distances under bone edges to monitor more distant sites. For example, this type of electrode is often passed around the temporal lobe to monitor inferior and mesial temporal structures.

Subdural grids consist of a rectangular or square array of contacts embedded in silicon (▶Fig. 3.2b). These constructs normally consist of 10 to 64 contacts; however, custom-designed grid patterns are available. These grids are placed on the cortical surface after a large craniotomy. An additional advantage to strip and grid electrodes is that these electrodes, unlike peg or other epidural electrodes, can be used for stimulation studies.

3.5 Planning Electrode Placement

Planning electrode placement requires the epilepsy surgery team to develop a patient-specific hypothesis regarding the possible locations of the patient’s epileptogenic focus (or foci). All components of the noninvasive evaluation are carefully considered to formulate this hypothesis. Invasive electrode investigation is pursued only when there are multiple possible locations of the patient’s epileptogenic focus (or foci), because patients with an evaluation that points to a single focus would undergo a resection without invasive evaluation.

One of the primary considerations is whether or not the patient’s epileptogenic focus has been lateralized. This generally relies heavily on the scalp EEG investigation in conjunction with seizure semiology. Interictal epileptiform activity, electrographic seizure-onset localization, and early spread of electrographic seizure activity are analyzed. The epileptogenic focus is considered lateralized when a patient has a stereotypical seizure semiology (particularly at onset) pointing to a single side of onset and EEG findings consistent with that side of onset. In such patients, a unilateral invasive electrode evaluation is pursued (see below). For patients with a nonlateralized epileptogenic focus, a bilateral electrode evaluation is pursued. Most of these patients have evidence suggesting a temporal epileptogenic focus, often felt likely to be medial temporal. Another subset of these patients is thought to have an extratemporal epileptogenic focus with rapid spread to the opposite hemisphere. These foci are often thought to be medial. In nearly all of these patients, there is a concern that they may harbor independent foci in both hemispheres. Various implantation approaches can be used for these patients, including epidural (typically peg electrodes), subdural strip, and depth electrodes. Generally, a common feature of all of these approaches is that electrodes should be symmetrically placed. This is done to minimize the chance of false lateralization due to rapid spread to the opposite hemisphere. Such contralateral spread typically occurs to the mirror image area. Having the same cortical areas in each hemisphere covered by electrode contacts is expected to detect seizure activity ipsilateral to the seizure onset prior to spread to the contralateral mirror region. The earliest bilateral invasive evaluations used peg electrodes for pinpointing lateralization. ²¹ Foramen ovale (FO) electrodes ²² , ²³ , ²⁴ were another early modality used for lateralizing seizure onsets in patients with presumed medial temporal focus (or foci). ²⁵ These were wires placed via needles inserted through the cheek and FO to reach the subarachnoid space adjacent to the uncus and medial temporal lobe. In the late 1980s and the early 1990s, a number of epilepsy centers used peg and FO electrodes extensively to lateralize seizure onsets predominantly in patients with evidence for temporal lobe seizure onsets. Also in the 1980s, as methods for the accurate stereotactic implantation of depth electrodes became more widely available, symmetrical placement of depth electrodes was increasingly adopted as the method of choice for lateralizing seizure onsets. ²⁶ , ²⁷ , ²⁸ , ²⁹ Depth electrodes are addressed in a separate chapter. Many epilepsy surgery teams have primarily relied on bilaterally placed electrode strips for determining seizure lateralization. ³⁰ , ³¹ , ³² , ³³ Often, more than four electrode strips are placed on each side to sample various brain areas. Strips are often placed in subtemporal, subfrontal, and interhemispheric locations. As noted earlier, brain areas covered are chosen based on the hypothesis generated from the noninvasive evaluation. Symmetrical placement is important, because contralateral spread occurs from one brain region to its corresponding contralateral counterpart.

For patients whose seizure onsets have been lateralized to one hemisphere, invasive evaluation is accomplished by implantation of multiple depth electrodes (referred to as SEEG), subdural grids and strips, or by a combination of depth and subdural electrodes. Many centers focus predominantly on one of these approaches, while others utilize all three of these options, making the decision based on the hypothesis for each individual patient. In most cases, the decision depends on the experience and preference of the epileptologist(s). This chapter is focused on the implantation of subdural grids and strips for these patients. Brain areas to be covered by the subdural electrodes are areas suspected to be part of the EZ and may also be areas of potentially eloquent functional cortex. This decision is based on analysis of all of the available data including seizure semiology; interictal and ictal scalp EEG findings; brain MRI, PET, and SPECT scans; neuropsychological evaluation; and IAP results. Seizure semiology often suggests the lobe of seizure onset. An aura of fear or déjà vu with progression to behavioral arrest, automatisms, and postictal nose wipe are typical of medial temporal lobe onset. Brief nocturnal hypermotor seizures often indicate a frontal lobe onset. Visual or somatosensory auras often indicate a parietal or occipital lobe seizure onset. Once the available data have been analyzed, the epilepsy surgery team composes a schematic diagram (▶Fig. 3.3) to indicate the desired placement of electrodes. As an example, if seizure onsets have been well localized to the patient’s dominant temporal lobe, but there is uncertainty as to whether the onset region is medial or lateral, electrode placement is planned to cover lateral, inferior, and medial temporal areas, including the posterosuperior temporal gyrus region (to enable mapping of language cortex, if needed).