Over the last two decades, advances in electroencephalography (EEG) and video-recording technology have fueled a dramatic evolution in the sophistication and number of epilepsy monitoring units (EMUs).^6,32 This chapter focuses on the specialized utility of such units. Other chapters in these volumes deal extensively with interictal EEG (Chapter 73), ictal EEG (Chapter 74), EEG mapping (Chapter 76), computer-assisted data collection and analysis (Chapter 95), the presurgical evaluation (Chapter 169), invasive monitoring (Chapters 170 and 171), and psychogenic seizures (Chapter 207). As a result, while these topics will be mentioned in the context of the EMU, reference should be made to other chapters for more detailed descriptions.

Previous guidelines for video-EEG monitoring have been published,⁸ but the most recent ones are those of the International League Against Epilepsy’s (ILAE) Commission on Diagnostic Methods Subcommittee on Neurophysiology for long-term monitoring.³³ The ILAE recommendations address monitoring both within and outside (e.g., intensive care unit [ICU]) the EMU (Table 1).

There are two main groups of patients who benefit from inpatient video-EEG monitoring: (a) those in whom the diagnosis is in question or needs to be confirmed and (b) those in whom epilepsy surgery is being considered. Epileptic seizures are episodic brief events, with each ictus typically lasting 2 minutes or less. While the history remains the most important component of the initial evaluation, there are many instances where the clinical details are ambiguous or insufficiently detailed to allow a definitive diagnosis. Patients rarely have seizures during outpatient visits or, with notable exceptions (e.g., unmedicated absence seizures), during routine EEGs. The history may be vague or uncertain because the patient has limited or no memory of the event, there were no observers, or the seizure occurred at night. Multiple types of episodes are sometimes reported, further confounding accurate diagnosis. Interictal EEGs may be normal or show only nonspecific abnormalities. Indeed, even with repeat EEGs about 30% of patients with epilepsy have normal routine EEGs. While a normal EEG does not exclude epilepsy, it also does not assist in making a diagnosis when the question of epilepsy is unresolved. Increased restrictions on administering conscious sedation have made it much more difficult to obtain routine outpatient sleep recordings, which can increase the yield of interictal abnormalities.

The EMU offers the benefits of prolonged recordings with video correlations in a controlled inpatient environment (Table 2). While the recording setup necessarily restricts a patient’s activity, this limitation is outweighed by the ability to reduce or withdraw medications if needed, something that is difficult to do safely in the outpatient setting. Simultaneous video recording greatly improves artifact recognition and reduces the likelihood that rhythmic or sharply contoured artifacts are misinterpreted as “epileptiform,” something that occurs more commonly with ambulatory recordings that do not include video. In addition to analysis of specific events, longer recordings improve assessment of interictal abnormalities. Medication changes can be implemented much more rapidly in the EMU than is possible in the outpatient clinic. In addition, one can more accurately determine the effects of medication. For example, patients who have their antiepileptic drugs (AEDs) discontinued may become aware of cognitive and other side effects that they had not previously recognized as drug related. Inpatient observation also allows trained professionals to interact and assess patients during seizures. This can provide valuable information regarding the precise nature of the ictal symptomatology and localization of the discharge. Also, if the EEG is screened by trained personnel (a frequent but not universal situation), inpatient monitoring may also help in differentiating and characterizing ambiguous EEG patterns. For example, interacting with children during bursts of diffuse rhythmic slow waves or slow spike-wave discharges permits alterations in responsiveness to be determined more accurately. Finally, inpatient monitoring allows alterations in recording methodology to be made “on the fly” depending on the EEG findings. This may be particularly helpful in suggesting the use of additional electrodes or the use of polygraphic techniques in selected patients.

Disadvantages of video-EEG monitoring include the need for hospitalization and the inherent limitations on the patient’s normal behavior. Antiepileptic drug withdrawal increases the risk of generalized tonic–clonic seizures and status epilepticus. Even with prolonged recording after AEDs have been withdrawn, events that occur infrequently may not be captured for analysis.