Long-term monitoring with CCTV and EEG may facilitate patient management by permitting more accurate classification of seizures and thus more appropriate therapy. For example, complex partial seizures can resemble the absence seizures of primary generalized epilepsy, but they are treated differently. Distinguishing between partial seizures with rapid secondary generalization and primarily generalized seizures also has important therapeutic implications.

Long-term monitoring can be used to distinguish epileptic seizures from intermittent nonepileptic behavioral alterations that resemble epilepsy. Although specific ictal behaviors suggestive of psychogenic seizures have been described,⁴⁴ they are absent in many, if not most, patients with psychogenic seizures.⁷¹ It may be impossible to distinguish between psychogenic and epileptic seizures on clinical grounds alone without LTM.

Because epileptiform discharges occur in people who do not have epilepsy,^24,48 including those who have psychogenic seizures,¹⁰⁵ LTM may be required to record seizures and confirm a diagnosis of epilepsy even if a routine EEG shows such epileptiform discharges. Psychogenic and epileptic seizures may coexist. Although some investigators have claimed that this is rare,⁷² others report that a substantial percentage^{20,65,93,105,132} or even a majority^56,109 of patients with confirmed psychogenic seizures also have, or have had, epileptic seizures. Thus, LTM may be necessary in a patient with known epilepsy if the current episodes are suspected of being nonepileptic. For example, if seizure frequency increases in such a patient despite an unchanged treatment regimen or if new episodes that differ from the previous seizures develop, the nature of the current episodes must be determined. Increased epileptic seizures might require increased doses of antiepileptic drugs (AEDs) or other therapeutic alterations. Increasing the AED dose in a patient with psychogenic seizures does not benefit the patient and may lead to drug toxicity.

Long-term monitoring may be used to quantify seizures such as subtle infantile spasms⁴³ or frequent brief absences that are not clinically obvious but interfere with a child’s school work.^50,69,120 It may also be used to assess the response to AEDs.^97,120 Quantification of seizures and measurement of blood AED levels when the AEDs are being administered by hospital personnel may be of value in cases in which patient compliance with the medication regimen and the validity of information provided about seizure frequency are both questionable.

Identification of IEDs is useful in the assessment of patients with epilepsy,^29,52,62,111 and it is also important in some patients who have not had seizures (Chapter 73). For example, the presence of IEDs in a child with language regression from a normal baseline but no history of seizures establishes a diagnosis of Landau-Kleffner syndrome.¹²

Interictal epileptiform discharges may be infrequent, and often they do not appear in routine EEG studies of patients who are known to have epilepsy. A single EEG demonstrates IEDs in approximately 30% of adults with partial seizure disorders.^2,112 The yield is higher when the patient is a child^2,22,138 or has a generalized seizure disorder.^121,138
Sleep and sleep deprivation also increase the yield of IEDs.^{22,41,84,115,138} Multiple routine EEG studies are of limited value, and may fail to demonstrate IEDs in as many as 41% of patients with partial epilepsies.¹¹² Of patients in whom IEDs are detected by routine EEG studies, the first EEG shows them in 45% to 81% of cases.^{112,114,121,138} Four EEGs identify 92% to 99% of patients in whom IEDs are ultimately found by routine recordings; each subsequent recording offers a minimal additional yield. Long-term monitoring covering the entire 24-hour day may demonstrate rare IEDs in patients in whom the results of repeated routine EEGs are negative.

Long-term monitoring is used to record seizures and localize their origin in patients with medically intractable seizures who are candidates for epilepsy surgery (see Chapter 169). Among epilepsy surgery centers responding to a 1991 survey,⁷⁶ 81% always used CCTV-EEG monitoring in the preoperative evaluation of patients with limbic epilepsy, 15% used it sometimes, and only 3% reported that they never used it. Multiple seizures must be recorded^103,128 to ensure that they all arise from the same focus or, if they do not, to quantitate the relative activities of the various seizure foci. Seizures that occur in clusters or flurries are more likely to arise from the same focus and thus carry less localizing information than seizures that are widely separated in time.⁵¹ A patient with partial epilepsy with multiple foci may still be considered a surgical candidate if the vast majority of the seizures come from one resectable site.¹¹⁷ Because seizures are only rarely captured during routine EEG, LTM is a necessary prelude to surgical treatment in most patients. Long-term monitoring using implanted (subdural and depth) electrodes may be necessary if the data from less invasive (e.g., sphenoidal) and scalp electrodes are not sufficiently localizing. In the 1991 survey, 92% of the epilepsy surgery centers reported that they used invasive recordings in at least some cases, although not necessarily in all patients.³⁹

Although unilateral IEDs may have substantial localizing value,⁶² examination of IEDs is generally not regarded as sufficient for localization of the seizure focus before surgery. In one study,¹³⁶ the lateralization suggested by interictal recordings alone was misleading in 8 of 67 patients with clear unilateral temporal lobe epilepsy. Moreover, many patients with good outcomes after temporal lobectomy have independent bitemporal IEDs on preoperative EEGs.^{29,117,123,136} Long-term monitoring shows a unilateral origin of the seizures or a strong lateral predominance in about half of the patients with bitemporal IEDs when only extracranial electrodes are used and in approximately 75% when depth electrodes are used.^54,116 The degree of lateralization of IEDs may have prognostic significance for patients undergoing temporal lobectomy^{10,29,45,52,104} even if LTM with depth electrodes shows unilateral seizure onsets.

Long-term monitoring is also used in neurological^53,60,96,130 and neonatal¹²⁴ intensive care units (ICUs) to identify deterioration of cerebral function, titrate drug doses for barbiturate coma, detect seizure activity in comatose patients, and confirm that treatment for seizures is successful. If a patient in status epilepticus has been treated with high doses of barbiturates, benzodiazepines, or other sedative or anesthetic drugs, continued obtundation could be a consequence of either medication effects or continued seizure activity. Only the EEG can distinguish between these and guide further treatment.

Seizures cannot be identified clinically in patients who have been pharmacologically paralyzed. When status epilepticus is prolonged, continued electrographic seizure discharges may not produce clinical seizure activity, even in patients who are not paralyzed.²⁵ Because status epilepticus can cause brain damage even in the absence of overt convulsions and metabolic derangements,^61,66,133 it must to be detected and treated. EEG monitoring demonstrates nonconvulsive seizures, including nonconvulsive status epilepticus, in a substantial percentage of neuro-ICU patients.^61,130

During LTM, surface electrodes are securely attached to the patient’s head, typically with collodion, to permit prolonged EEG recordings. Electrode impedances should be both low and balanced to minimize noise and maximize the effective common mode rejection ratio of the recording system.⁷⁰ Open-hole cup electrodes are used for recordings lasting days to weeks, so that the electrodes can be periodically regelled. The quality of the recordings must be continually assessed so that electrode maintenance can be performed as soon as possible if the data deteriorate.

The patient’s head is wrapped to protect the electrodes and to prevent infection when implanted electrodes are used. Strain relief loops to secure electrode lead wires are especially important when implanted electrodes are used. In patients who become confused or combative during or after seizures and who might pull on the lead wires of their implanted electrodes, judicious use of restraints may be necessary to prevent self-injury.

Multichannel EEG data are amplified, undergo initial bandpass filtering, and are recorded on a storage medium, most commonly magnetic tape. Guidelines for amplifier specifications for LTM have been published.^7,37 If the EEG signal is digitized, the analog-to-digital converter must satisfy the Nyquist criterion: The sampling rate for each input channel must be at least twice the highest frequency present in the analog data. Long stretches of cable and wire will pick up electromagnetic and movement artifacts. The higher the amplitude of the EEG signals traversing these wires, the better is the signal-to-noise ratio. Therefore, preamplifiers placed close to the recording electrodes, either on the patient’s head (within the head wrapping) or in an electronics package attached to the patient with a strap or belt, are preferable to remote preamplifiers. When the EEG signal is carried from the patient to the recording machinery by cables (cable telemetry), careful attention must be given to patient isolation to prevent electrical hazards.

If radio telemetry is used, the data must be multiplexed; sample values from all the input channels are combined into a single radio signal. Multiplexing is also used for storage media in which a single “channel” contains all the EEG data, as in videotape recording systems; successive data values stored on the tape are derived from consecutive input channels, going from the first to the last channel and then back to the first channel again. The maximum frequency of the data stored for each channel, which must also satisfy the Nyquist criterion, is equal to the storage frequency (samples per second) of the tape divided by the number of channels that are multiplexed together.

The EEG may be recorded either in a referential format, with the same electrode connected to the input 2 of each channel’s differential amplifier, or in a bipolar montage. The referential recording permits reformatting of the retrieved EEG data to any desired montage.⁶⁷ This requires carefully balanced amplifiers and high-fidelity recording and playback so that the calculated bipolar data will faithfully reproduce the signal that a bipolar
analog recording would have provided. A faulty reference electrode can make all of the recorded patient data uninterpretable. Thus, referential recordings are usually confined to a hospital setting (inpatient LTM or daytime intensive monitoring), in which the quality of the data can be assessed frequently and technical problems rapidly fixed. Additional recording channels may be used to record the electrocardiogram, electromyogram, respiratory activity, and other physiologic variables, as well as alarm or marker signals that can be triggered by the patient or hospital staff.

The EEG can be analyzed in real time by a computer program that identifies and marks probable epileptiform patterns (computerized EEG analysis is discussed in Chapter 95) for subsequent analysis. In an inpatient setting, the computer can alert the staff so that the patient’s responsiveness, neurologic status, and memory can be tested during a seizure.

The EEG data were formerly stored on magnetic tape, but increases in the speed and capacity of computer disks have made them the preferred storage medium. Data storage in a random-access memory device such as a computer disk also has the advantage that any portion of it can be accessed quickly; with magnetic tape, winding to the desired part of the record can markedly prolong data access times.

When CCTV and EEG data are recorded concurrently, all data should include time markers⁵⁷ or other methods that permit them to be correlated or synchronized so that the temporal relationship between EEG changes and behavioral alterations can be examined. Many different kinds of CCTV instrumentation are available.^7,37 Cameras sensitive to low light levels or infrared illumination are useful for monitoring patients at night. Multiple cameras with remote-controlled pan and zoom permit recording of a split-screen image that contains both a close-up of the face and eyes and a view of the patient’s entire body and that tracks patients’ movements to keep them in view as much as possible.

Long-term monitoring procedures can be subdivided according to whether they are performed on an outpatient or inpatient basis and by the specific technology utilized. In outpatient LTM, the patient carries an electronics package and recording device. Initially the EEG was stored on magnetic cassette tape, but advances in semiconductor memory technology now permit ambulatory EEG (AEEG) recordings utilizing solid-state memory arrays. Inpatient LTM involves a hospital admission to a specially equipped monitoring unit in which the patient is monitored continuously for days or weeks. In an intermediate approach, which has been labeled daytime intensive monitoring,¹¹⁰ the patient undergoes repeated prolonged EEG or CCTV-EEG recordings in the hospital but is not admitted to the hospital and goes home each night. The patient is encouraged to sleep during the recording session; antecedent sleep deprivation may facilitate this. When monitoring is performed in the hospital, either on an inpatient basis or as daytime intensive monitoring, the patient can be connected to the recording equipment with a long cable (an arrangement sometimes referred to as cable telemetry) or the data can be sent by radio or infrared telemetry from a transmitter worn by the patient to a nearby receiver and thence to the recording equipment. During cable telemetry, the same semiconductor memory technology used for AEEG can permit continued EEG recording for a limited time when the cable is disconnected, such as when the patient goes for a procedure in another part of the hospital.

Each of the LTM methods has advantages and disadvantages. For example, considerations of equipment weight and size, recording medium capacity, and battery life place limitations on the number of channels and the duration of recordings that can be obtained during AEEG monitoring, although both have been increased as the technology evolves. The decision about which method to use should include consideration of the particular clinical situation of each individual patient.