FUTURE DIRECTIONS IN AMBULATORY EEG
STEVEN C. SCHACHTER, MD
There are numerous current uses for ambulatory EEG (aEEG) monitoring as detailed in other chapters, including confirmation of a clinical suspicion of epilepsy, documentation of seizure type(s) and frequency by recording in an environment where a patient has the events of interest, and screening for possible nonepileptic seizures, including movement disorders and psychogenic nonepileptic seizures (1,2).
With advances in aEEG monitoring equipment, technical issues producing an uninterpretable recording have been reduced. Nonetheless, fixing technical problems that arise in the field and identifying the nature and source(s) of artifacts and their differentiation from electrographic seizures are still problematic. In addition, assessment of very infrequent events is not possible with current technology, tapering antiseizure drugs (ASDs) to evoke seizures presents safety risks, and patients or family members may not maintain the seizure diary as completely and accurately as they do in the epilepsy monitoring unit (EMU). Another disadvantage is that certain seizure types may not be reliably detected or distinguished with use of only EEG, but even when video recording is used in conjunction with EEG, patients may be off camera or under bed sheets during events, thereby limiting the diagnostic value of the video component (3).
Unmet Needs in Monitoring Patients With Epilepsy
Given these and other limitations, aEEG technology is primarily used for short-term monitoring to facilitate diagnosis. Yet, seizures are unpredictable and have associated risks of fractures, intracranial hematomas, burns, accidents, aspiration, drowning, and sudden unexpected death (SUDEP); high rates of drug-resistant seizures continue despite availability of new ASDs and cognitive and other issues negatively impact on treatment adherence. Therefore, the quality of life (QOL) of patients with epilepsy as well as the ability of physicians to care for patients and for patients to be self-empowered could substantially be enhanced by new systems that incorporate EEG and/or other modalities that could reliably and consistently:
• Provide patients or caregivers with a warning of an impending debilitating seizure with sufficient time to protect the patient’s safety
• Couple seizure detection to a therapeutic intervention (closed-loop)
• Enhance safety and lessen risk of injuries during and after a seizure
• Summon help or emergency evaluation when needed but not otherwise
• Identify and monitor biomarker(s) for SUDEP
• Measure the impact of treatments on seizure control and comorbidities
• Improve adherence to medication and lifestyle issues (eg, sleep)
• Assess quality and adequacy of sleep and comorbidities, such as anxiety and depression
• Correlate specific symptoms with EEG, ASD levels, and other physiological signals in close to real time
• Supplement patient diaries with objective records of seizure type, frequency, and severity for use in the clinic and during clinical trials of antiseizure therapies
Given these opportunities, aEEG monitoring is evolving from EEG- and video-based devices for short-term diagnostic use to long-term comprehensive epilepsy management systems incorporating a range of technologies to improve the QOL of persons with epilepsy. Many such enabling technologies, whether based on EEG or other modalities, are emerging either de novo for use in epilepsy monitoring or by adapting technologies that were initially developed for other applications. A representative sample of the technologies, especially those that have been presented in public forums (eg, the Epilepsy Foundation Pipeline Update Conference (4)), is presented in the following section. Technical discussions related to development of seizure detection/prediction algorithms and other engineering design aspects are beyond the scope of this chapter.
EMERGING TECHNOLOGIES AND POTENTIAL APPLICATIONS
Epilepsy monitoring technologies may be categorized by whether they are EEG- or non-EEG-based, and by whether their intended use is primarily for ambulatory seizure detection/prediction (see (5–7) for reviews) and therefore to facilitate therapy (8) or prevent injuries/death or to self-empower persons to live with their epilepsy. The rapid pace of technological development, for example, compared to drug development, suggests that the availability, capabilities, and features of systems described in the following will likely change over time. However, the unmet needs they seek to address will remain until practical, affordable, and reliable solutions are developed, tested, validated, and implemented.
Ambulatory Seizure Detection/Prediction Systems
EEG-Based Systems: Extracranial
Many scalp-EEG electrode systems have been designed for gaming or for brain–computer interfaces and serve a variety of uses from entertainment to research to cognitive remediation (9–11). However, they are generally not practical for chronic epilepsy monitoring because they are too conspicuous and prone to slippage, noise, and other artifacts, especially due to poor electrode–scalp conductivity, which has prompted efforts to identify adhesives that work well for a watertight electrode–scalp interface (4,12). Therefore, electrodes that are inconspicuous, easy to apply, and less subject to slippage on the scalp are preferable. For example, Epitel, Inc. is developing a wireless, waterproof, disposable EEG patch for scalp placement to document occurrence and duration of seizures using flexible circuit boards and two gold electrodes that are positioned based on a patient’s prior EEG findings (4). The device is covered by flexible urethane and once applied to the scalp starts to log and transmit EEG data immediately and for a period of 3 days, though with further development the recording time will be extended to 7 days. Multiple devices could be placed on the scalp if desired.
Other groups are developing wireless electrodes that would be inserted under the scalp using small incisions and therefore be invisible to others, including one system that is designed to record and store EEG using a behind-the-ear platform (Figure 10.1) (13).
EEG-Based Systems: Intracranial
Intracranial EEG-based systems for chronic seizure monitoring, while more invasive than extracranial systems, are further ahead in development. The Responsive Neurostimulation System™ (Neuropace, Inc.) delivers closed-loop stimulation in response to detected epileptiform activity and received Food and Drug Administration (FDA) approval for use in adults with partial-onset seizures not controlled with greater than or equal to 2 ASDs. The device consists of a cranially implanted, programmable, battery powered, microprocessor-controlled neurostimulator that is connected to depth and/or subdural strip leads. The system appears to be well tolerated and self-reported median seizure reductions at 1 and 2 years were 44% and 53%, respectively (14). The Seizure Advisory System (originally developed by NeuroVista) utilizes intracranial electrodes to transmit ongoing EEG to a subclavicular, implanted telemetry unit. Data are then sent wirelessly to an external, handheld personal advisory device that uses a patient-specific seizure prediction algorithm to display indicator lights for seizure likelihood—blue (low), white (moderate), or red (high). A published study of 15 ambulatory patients with 2 to 12 disabling partial-onset seizures per month showed that intracranial EEG monitoring for seizure prediction is feasible (15). The enrolled patients entered a data collection phase, during which an algorithm for identification of periods of high, moderate, and low seizure likelihood was established for each patient. Eleven out of 15 patients had likelihood performance estimate sensitivities ranging from 65% to 100% and entered an advisory phase. Clinical effectiveness measures remained stable between baseline and 4 months after implantation.
Non-EEG-Based Systems: Accelerometers
Non-EEG-based seizure detection systems take advantage of changes in physiological signals other than EEG that stereotypically change with seizures, such as limb movements in generalized tonic–clonic seizures (GTCSs). Accelerometers detect changes in velocity and direction, and some are designed to determine movement in the x, y, and z planes. The SmartWatchTM (Smart Monitor) consists of a Global Positioning System (GPS) module and a proprietary accelerometer/gyroscopic sensor to detect the excessive and repeated motions of GTCSs, in addition to keeping time (4). Algorithms continuously monitor and analyze wrist motion to detect repetitive shaking movements from a GTCS and then automatically send a text message and/or phone call to caregivers or other designated alert recipients, providing the location of the patient based on GPS. A button on the watch can be pressed by the patient in case of an emergency, if they feel that they are about to have a seizure, when a non-GTCS (such as a complex partial seizure) has occurred, when medication was taken, or in the event of a false detection. The SmartWatch also provides medication reminders, analyzes sleep duration and quality, records audio during seizure episodes, and provides analytics and reporting/seizure tracking for physicians, including seizure duration, severity (amplitude and frequency of the shaking movements), frequency, time of occurrence, and associated audio.
In one published study of the SmartWatch, patients aged 6 to 68 years were monitored in an EMU for 17 to 171 hours (total of 4,878 hours) (16