Monitoring epilepsy patients with scalp-recorded electroencephalography (EEG) is the mainstay in the evaluation for potential curative surgery. Inpatient combined video-EEG monitoring (VEM) to record seizures can be used to help delineate the general region of the brain in which seizures start in a specific patient, as well as determine the type of seizure and exclude the diagnosis of nonepileptic seizures. In addition to the information gained by EEG recording, other noninvasive tests can be used to determine the ictal onset zone and underlying pathology. However, in a select group of patients, intracranial EEG monitoring is needed for determining eligibility for epilepsy surgery and for planning that surgery. The evaluations of these patients identify whether a resectable focus is present when noninvasive methods have not yielded the concordant findings needed to proceed to resective surgery.

Intracranial VEM has no role with patients who have been excluded as candidates for surgery. This could be for as simple a reason as patient choice. Some patients would not consider brain surgery of any kind, either resection of brain tissue or implantation of a stimulation device, and would therefore not be appropriate for intracranial VEM, given that there are associated risks and that the study could provide no useful information for these patients. Other patients can be excluded as candidates for epilepsy surgery based on scalp EEG and other noninvasive diagnostic modalities. Such patients include those who clearly have multiple ictal onset zones or have generalized onset seizures. Some of these patients may be candidates for corpus callosotomy, but this surgery does not require intracranial VEM.

Other patients are clearly candidates for epilepsy surgery, and this determination and the planning for the surgery can be completed using noninvasive techniques. High-resolution magnetic resonance imaging (MRI) has been the most effective scanning technology in expanding the group of surgery candidates who do not require intracranial VEM. Patients with focal abnormalities, such as tumors, vascular abnormalities, and hippocampal sclerosis, and have corresponding scalp EEG recordings demonstrating ictal onset in the region of the structural lesion may be able to proceed to surgery without electrocorticography based on the anatomical localization from the MRI.¹ Other imaging modalities, including positron emission tomography (PET), magnetoencephalography (MEG), and single-photon emission computed tomography (SPECT) sometimes are used to confirm the localization. Since the newer noninvasive technologies have become available, fewer patients who are potential candidates for epilepsy surgery now require intracranial electrode recording.

There are several types of intracranial electrodes. Seldom-used epidural peg electrodes provide more focused information about the region of seizure onset than scalp electrodes, but they do not provide the spatial resolution necessary for surgical intervention and very rarely add sufficient information to justify their use over noninvasive scalp EEG (Figure 30-1). The two types of intracranial electrodes most commonly used for preoperative evaluation are subdural and depth electrodes. Subdural electrodes come in multiple arrangements, from strip electrodes that can be inserted through bur holes or under the edges of a craniotomy to grid electrodes of various sizes that cover an entire brain region under investigation. Depth electrodes are stereotactically implanted into the parenchyma of the brain and can be used alone or in combination with subdural electrodes. Our discussion will focus on depth electrodes and their use, based on the clinical experience of the epilepsy surgery program at the University of California at Los Angeles (UCLA).

Depth electrodes are most effectively used when direct recording of structures beneath the cerebral surface is needed. The most common deep structures recorded with depth electrodes are the medial temporal structures, including the hippocampus and amygdala. Frequently, preoperative evaluation of medial temporal lobe epilepsy can be accomplished with noninvasive techniques. If the patient has an MRI demonstrating hippocampal sclerosis, concordant scalp ictal EEG showing seizure onset in the same temporal lobe, and at least one additional corresponding test (PET, MEG, SPECT, or neuropsychological testing), many centers will proceed with anteromedial temporal lobectomy if the epileptogenic hippocampus does not support memory. However, if these tests are not localizing or somehow contradict each other, depth electrodes can provide direct electrographic information from the hippocampus with optimal spatial and temporal resolution for deep structures.

Some epilepsy centers use subtemporal subdural strip electrodes placed through bur holes to record from the hippocampus, seeking to avoid the risk of parenchymal hemorrhage from depth electrodes.^2,3 Studies comparing hippocampal seizure activity from depth electrodes and subdural strip electrodes in patients with both types of electrodes placed bilaterally have had mixed results. Two early studies found that subdural electrodes were less sensitive than depth electrodes, but they disagreed on the efficacy of subdural electrodes in lateralizing the seizure onset, with one study finding errors in lateralization,⁴ and the other finding that subdural electrodes were consistently accurate in lateralizing temporal lobe epilepsy.⁵ Two commentaries in response to the study finding lateralization errors also disagreed as to the accuracy of subdural electrodes for lateralizing temporal lobe epilepsy.^6,7 Depth and subdural electrodes appear equivalent in detecting interictal spike activity from the medial temporal lobe.⁸ Errors in ictal onset lateralization, however, have been reported with subdural electrodes,⁹ and more than one article has argued that depth electrodes are better for determining seizure onset within the medial temporal lobe.^10,11 Because there are no definitive answers, hippocampal depth electrodes would seem to be the most reliable for lateralizing seizure onset in medial temporal lobe epilepsy patients. However, for patients thought to have dual pathology involving hippocampal sclerosis with a separate potential neocortical focus as well, combined subdural and depth electrodes appear to be the appropriate choice to address both pathologies,¹¹ although orthogonal depth electrode placement enables sampling along the electrode shaft of both neocortical and medial temporal sites simultaneously.

Depth electrodes are long, thin electrodes with multiple contacts exposed along their length designed to record electrical activity at several increasing distances from the electrode’s hilt (Figure 30-2). Typically, 4 to 18 contacts are spaced every 5 to 10 mm along the length. The electrodes are most commonly made of platinum-iridium, but they have been constructed of other alloys.^12,13 Both rigid and flexible depth electrodes have been used, although most centers are currently using flexible electrodes. The flexibility of depth electrodes results in a lower risk of brain injury and hemorrhage. Their disadvantages are the difficulty in guiding a flexible electrode exactly to the target region and their propensity to withdraw during the placement and securing process. Insertion cannulae and stylets, as well as guide screws with sealing caps, have been effectively used to address these problems.

Although freehand placement of depth electrodes has been reported with some degree of accuracy,¹⁴ depth electrodes are usually placed with either stereotactic frame-based or stereotactic frameless systems with the patient either awake or under general anesthetic. Computed tomography (CT)¹⁵ or MRI¹⁶ can be used for the stereotactic planning (Figure 30-3). The use of the frameless stereotactic system for accurate placement of depth electrodes has been reported in the literature.^17,18 Multiplanar and probe-oriented views may improve some aspects of depth electrode placement and can be used in the planning of frame-based depth electrode placement.¹⁹ Some of the available frameless stereotactic systems have locking arm mechanisms to aid in the placement of depth electrodes. The advantage of frameless stereotactic systems is that they make combined studies using both subdural and depth electrodes placed via craniotomy easier than with a frame-based stereotactic system. Nevertheless, we prefer the frame-based system because of the greater potential accuracy in electrode placement. The major commercially available stereotactic frames are the Cosman-Robert-Wells (CRW) frame marketed by Integra Radionics Inc. (Burlington, Massachusetts) and the Leksell frame from Elekta AB (Stockholm, Sweden). Both frames are effective for very accurate placement of depth electrodes, with specific strengths and weaknesses of each. The advantage of the CRW frame is that the nonsterile portion attached to the patient’s head can be completely draped off, eliminating the risk of inadvertent contamination as the coordinates are changed. This is especially important when the number of electrodes placed requires the frame coordinates to be adjusted several times. Also, the arc of the CRW system can easily be shifted from the coronal to the sagittal plane, which is necessary for orthogonal placement of electrodes, as will be discussed later. The CRW frame has been favorably reviewed for depth electrode placement in the literature.²⁰ At UCLA, the Leksell Model G frame is used with a special adapter for the Y-Z coordinate arc attachment ring (Figure 30-4). The vertical and anteroposterior coordinates are set, and the lateral coordinates are determined by the depth of electrode insertion. With this technique, all electrodes are placed orthogonally (perpendicular to the sagittal plane).

Both sagittal and orthogonal approaches have been described for placement of depth electrodes. The sagittal approach developed by Dennis Spencer, M.D. at Yale University was designed for recording activity from the entire extent of the medial temporal structures, including the length of the hippocampus and the amygdala.^5,13 With this approach, the depth electrodes enter from the occipital lobe and extend down the length of the hippocampus to end in the amygdala. Because these electrodes extend for a long distance, frame-based stereotaxy and an insertion cannula are necessary to ensure accuracy. The advantage of this approach is that the entire hippocampus and the amygdala can be sampled with a single electrode. There is some evidence of variability in location of seizure onset along the length of the hippocampus,²¹ and different surgical outcomes have been reported depending on the location of seizure onset along the length of the hippocampus,²² indicating that longitudinal depth electrode delineation of the exact seizure onset location within the hippocampus could be of clinical importance. However, several studies have shown that the more hippocampal tissue is removed, the greater the rate of seizure control in patients with medial temporal lobe epilepsy,^23,24 so evidence of an anterior or a posterior hippocampal seizure onset is unlikely to guide the surgeon to do a more limited resection. For this and other reasons, despite developing the longitudinal depth electrode approach, the Yale epilepsy surgery program no longer routinely uses it in the evaluation of epilepsy patients. It continues to be used by other epilepsy surgery programs.¹⁵