Utility of Stereoelectroencephalography in the Treatment of Drug-Resistant Epilepsy

A person with epilepsy is considered to have drug-resistant epilepsy (DRE) after adequate trials of 2 well-chosen antiseizure medications (ASMs) fail to achieve seizure freedom. DRE, which affects about a third of people with epilepsy, can be effectively treated with epilepsy surgery, resulting in seizure freedom in approximately 35% to 80% of appropriately selected patients. Neurosurgical options have proliferated in recent years with the addition of minimally invasive procedures, such as MRI-guided laser interstitial thermal therapy and neuromodulation approaches, such as responsive neurostimulation.¹ This has expanded the eligibility for neurosurgical intervention in DRE including both definitive and palliative procedures. Unfortunately, underreferral and delayed referral persist, affecting millions of people with epilepsy. Therefore, it is imperative for neurologists to closely monitor people with epilepsy for drug resistance, recognize surgically remediable syndromes, and refer such patients for presurgical evaluation in a timely manner.² Neurologists must also emphasize to patients and their families that epilepsy surgery should not be regarded as a treatment of last resort, but rather as a powerful management tool—alongside ASMs, neurostimulators, and dietary therapies—that can positively impact seizure frequency, as well as quality of life.

Multidisciplinary Presurgical Evaluation
The success of epilepsy surgery depends on a carefully designed, personalized, neurosurgical hypothesis and a plan to test the hypothesis. Although patients with epilepsy caused by brain tumors or other lesions can sometimes be operated on to attain seizure freedom, a majority of patients—including those who are MRI-negative or have complex or multifocal lesions—require a stepwise evaluation to define the epileptogenic zones (EZs), localize functionally eloquent cortices, and optimize the surgical target(s). This evaluation is typically facilitated at level 4 National Association of Epilepsy Centers with multidisciplinary expertise including epileptologists, neurosurgeons, neuropsychologists, neuroradiologists, neuroscience-trained nurses, care managers, and other supportive and administrative staff.

Presurgical Evaluation Steps
After a detailed history and neurologic examination, the first (noninvasive) phase of presurgical evaluation involves a combination of the following tests: brain MRI, scalp-based video EEG, PET, single-photon emission CT, magnetoencephalography (MEG), and neuropsychological evaluation. In some cases, functional testing is also performed, including functional MRI, transcranial magnetic stimulation, functional MEG, EEG source imaging (ESI), and Wada test.³ These tests are individualized for each patient and aim to evaluate different aspects of brain function and dysfunction, including neuroanatomy (MRI), cortical excitability (video EEG [vEEG], MEG, and ESI), regional metabolism and perfusion (PET and SPECT), functional deficits (neuropsychology, Wada), and functional activation (fMRI, MEG, transcranial magnetic stimulation [TMS], and ESI). The results from these tests are synthesized to formulate one or more anatomic–electroclinical hypotheses regarding the location and geometry of the EZ and its relationship to functional cortices (Figure 1).⁴ In this article, EZ is conceptualized as the cortical region involved in the generation and primary organization of the seizures.⁴

Figure 1. Illustrative example of a noninvasive (phase 1) presurgical evaluation. Interpretation and synthesis of these findings by the multidisciplinary epilepsy surgery team form the basis for the hypothesis guiding the stereoelectroencephalography implantation scheme. All tests except single-photon emission CT (SPECT)/subtraction ictal SPECT coregistered to MRI (SISCOM) are from a boy aged 6 years with drug-resistant epilepsy due to tuberous sclerosis complex. Not all patients require every test. In addition to these modalities, functional testing may be needed, including functional MRI, transcranial magnetic stimulation, functional magnetoencephalography (MEG), and the Wada test. 

The second (invasive) phase is then undertaken to test the anatomic-electroclinical hypotheses generated by the first phase and involves intracranial EEG (iEEG). Two primary modalities for iEEG include subdural electrodes (SDEs) and stereoelectroencephalography (SEEG).⁵ Over the last 2 decades, SEEG has become the preferred iEEG modality worldwide due to its comparable effectiveness in localizing the EZ and functional cortices, superior safety and tolerability, ability to access deeper structures, and relative ease of bilateral sampling.^6,7

Individualized Stereoelectroencephalography Planning
The multidisciplinary analysis of a patient’s phase 1 data helps personalize an SEEG plan. Each electrode entails risk of complications. These include brain hemorrhages, infection, and mistargeting requiring reimplantation. Electrode placement can be consequential in certain brain areas; therefore, a SEEG implantation scheme is planned carefully with each electrode requiring specific justification. In general, sampled brain regions include abnormalities identified on noninvasive phase 1 testing or are chosen on the basis of seizure semiology, knowledge of brain network connectivity, patterns conserved across electroclinical syndromes, desire to triangulate the potential EZ, and anticipated functional brain mapping. Typically, 10 to 20 electrodes are implanted with each electrode having between 4 and 16 cylindrical contacts. Exact specifications may vary, but each contact usually has a diameter between 0.8 and 1 mm, a contact length of 2 mm, and a center-to-center distance 3.5 mm.⁸ EEG technologists with specialized training and experience in iEEG monitoring are crucial for troubleshooting technical issues that may arise after implantation and should be available throughout the patient’s admission.

Bedside Care of Patients Undergoing Stereoelectroencephalography
Following SEEG implantation, patients are often admitted to epilepsy monitoring units (EMUs). In EMUs, patients are managed by neurologists or neurohospitalists working alongside epileptologists. Hence, it is important for the team to understand the specific considerations involved in the care of patients receiving SEEG. Some of the early decisions include withdrawal or modification of the ASM regimen, formulating a seizure rescue plan, and educating caregivers and families to identify habitual seizures by pressing an event button, which serves 2 purposes: it marks the time on the SEEG tracing so that the event can be analyzed and also notifies the nursing team to provide appropriate assessment, including bedside cognitive and language testing, enabling this information to be relayed to physicians to ensure appropriate seizure management. 

After SEEG implantation, anesthesia typically wears off within the first 12 to 24 hours. Hence, in the first few days following implantation, patients may experience pain at the craniotomy sites of electrode placement, particularly with electrodes that enter the temporalis muscles. Although hemorrhagic complications are rare and are often asymptomatic, they usually depend on the number, size, and locations of electrodes. Patients receiving SEEG should be periodically evaluated for signs that might portend a complication. The clinical signs potentially reflecting complications include altered mental status, pupillary and extraocular motion abnormalities, surges in blood pressure, wide fluctuations in heart rate, and unusual focal iEEG changes on the SEEG.⁹ In well-resourced modern neurosurgical practice, infectious complications from SEEG placement are extremely rare; nevertheless, vigilance for cerebrospinal fluid leaks is desirable.⁹ In addition to these SEEG-specific complications, some patients may be susceptible to deep vein thrombosis or pressure sores from the prolonged period of relative immobility required for monitoring. Therefore, having an institutional protocol for anticipating and addressing such medical issues is important to optimize care. 

Patients receiving SEEG typically remain admitted until several habitual seizures are captured and functional brain mapping is completed. The duration of hospital admission is unpredictable and dependent on seizure occurrence, although most patients are monitored for ≤7 days. The whole process can be frightening and sometimes frustrating for patients and their families. Therefore, in addition to presurgical counseling, it is important to provide ongoing emotional and psychological support to patients during hospitalization. Adjunctive professionals, including Child Life specialists (for pediatric patients), integrative care providers, and social workers, can help make the EMU stay more comfortable. If feasible, patients can also be encouraged to keep up with schoolwork or employment during monitoring with support from their school or coworkers, which can help normalize the experience of being hospitalized. 

Stereoelectroencephalography Interpretation for Epileptogenic Zone Identification
Habitual seizures captured during SEEG monitoring are analyzed to determine the location and 3-dimensional extent of the EZ. This requires identifying and localizing the earliest sustained ictal change on the SEEG. Ideally, this electrographic change precedes the first definite clinical change, otherwise one must consider the possibility that the seizure onset zone has not been sampled. In routine clinical practice, the ictal EEG onset is analyzed visually, ideally using multiple examples of the habitual seizures, to determine the most consistent localization at onset. Many seizures exhibit characteristic onset patterns on SEEG, including low-voltage fast activity or low-frequency periodic spikes, although several other seizure onset patterns may occur.¹⁰ Some centers also use quantitative methods for ascertaining the EZ (eg, epileptogenicity index, connectivity metrics), although these are often ancillary and are not yet the standard of care.¹¹

In addition to monitoring for spontaneous seizures, many centers also use electrical stimulation to induce seizures, a method which has been shown to help localization and has prognostic value for outcomes of epilepsy surgery.¹² Interictal epileptiform activity is also analyzed carefully to identify regions of consistent abnormalities; for example, near-continuous, low-frequency, high-amplitude spikes intermixed with high-frequency activity on iEEG can sometimes indicate underlying cortical dysplasia.¹³

It must be recognized that SEEG is subject to a sampling bias due to its high-density sampling of widely separated brain regions, causing vast areas of the brain to remain unsampled.This is why SEEG must always be hypothesis-driven and should never be used as an exploratory modality (Figure 2). 

Figure 2. Invasive (phase 2) epilepsy presurgical evaluation with stereoelectroencephalography (SEEG). The first panel (A) shows a reconstruction of the SEEG implant obtained by digitally fusing cortical surface segmented from the patient’s brain MRI and electrode locations from the postoperative CT scan. The implant shown here is bilateral and asymmetric, with 10 electrodes in the left hemisphere and 3 electrodes in the right hemisphere, respectively. The next panel (B) shows the SEEG tracing with contacts arranged in a bipolar montage along the respective electrodes and depicts interictal epileptiform discharges. Although SEEG nomenclatures vary, one convention is to label left hemisphere electrodes with an apostrophe (eg, N′). Note the high-amplitude tonic discharges from the N′ and Li′ electrodes, typical for dysplasia and tubers. The bottom panel (D) shows an overview of ictal onset and propagation with a box highlighting the area of seizure onset, also shown in the panel above (C). The first sustained change is low-voltage beta to gamma frequency activity in a subset of contacts on the Li′ and O′ electrodes, both located in the left occipital lobe. These data were obtained from the same patient shown in Figure 1.

Stereoelectroencephalography for Identifying Functional Brain Regions
The neurosurgical plan in DRE must satisfy 2 requirements: critically interrupt the epileptic network and avoid or at least minimize any new neurologic deficits/functional consequences. This entails identifying functional cortices, including sensorimotor and speech/language areas and their relationship to the EZ.¹⁴ Previously, there have been concerns that SEEG may not be able to localize functional brain regions as reliably as SDE. However, this viewpoint has been challenged. SEEG compares favorably with SDE for localizing motor and speech/language regions and is superior to SDE for identifying sensory response.¹⁵ Functional localization with SEEG is routinely performed using electrical stimulation mapping (ESM). This involves attempting to elicit reproducible behavioral responses either passively (for sensorimotor mapping) or during a repetitive task (for speech/language mapping) when applying safe amounts of electrical current into the implanted electrodes.¹⁴

Special Considerations in Children and Adolescents
The core principles of SEEG planning, implantation, and interpretation are similar regardless of age. However, a few pediatric considerations are worth highlighting. From a neurosurgical perspective, most teams have their own criteria for minimum skull thickness (typically 2-3 mm) to support the bolts for SEEG electrodes and for patient’s weight (typically 10-12 kg) to tolerate the surgical procedure. In young patients (typically aged <2 years), the risks and benefits of SEEG monitoring require consideration with more than usual care. Some pediatric patients with behavioral and/or developmental challenges may be monitored safely and effectively under dexmedetomidine infusion in an intensive care unit. This helps ensure that the electrodes are not pulled out by the patient and that the data captured are of good quality devoid of movement artifact. In our experience, such patients often have high seizure burden and usually require only a short period of monitoring. 

Children are also relatively more likely to have multifocal epilepsies compared with adults, which can require more extensive or bilateral implantations. High seizure burden, often seen in such patients, also requires customization of the ASM plans. Pediatric patients, particularly those with epileptogenic pathology occurring early in life, are uniquely susceptible to functional cortical reorganization. While this necessitates performing ESM in such patients to delineate functional cortices, it may be challenging given the limited ability in some of these patients to sustain full cooperation during ESM. Emerging methods relying on task-related changes in SEEG power spectra and their correlation with observed behavior may improve the access of ESM to pediatric patients and provide new insights into cognitive neurophysiology.

The Exciting Future of Stereoelectroencephalography
At present, SEEG represents the standard of care to define the EZ and functional cortices in patients with complex DRE being evaluated for neurosurgical intervention. In addition, SEEG offers a unique opportunity to sample electrical activity of the neural ensembles at an unparalleled spatiotemporal resolution. Given the rapidity of the development and application of artificial intelligence in medicine and progress in biosensors, transfer learning approaches may allow noninvasive modalities to reliably estimate intracranial brain networks. As a result, in the foreseeable future, SEEG may no longer be required for presurgical evaluation in a subset of patients with DRE. Until such time, the field continues to expand and advance on the clinical front to refine our approach for evaluating patients with DRE using SEEG and on the scientific front as a tool to investigate the electrophysiology of the brain.