Ambulatory EEG

Updated: Nov 12, 2019
  • Author: Elizabeth J Waterhouse, MD, FAAN; Chief Editor: Selim R Benbadis, MD  more...
  • Print



Ambulatory electroencephalography (AEEG) monitoring is a relatively recent technology that allows prolonged electroencephalographic (EEG) recording in the home setting. [1] Its ability to record continuously for up to 72 hours increases the chance of recording an ictal event or interictal epileptiform discharges. AEEG is a less expensive alternative to inpatient monitoring, with costs that are 51-65% lower than a 24-hour inpatient admission for video/EEG monitoring. [2]

Continuous cardiac monitoring was first described by Holter in 1961. [3] The development of portable EEG recording proved more problematic than the Holter monitor because of the need for signal amplification and multichannel recording. A multichannel portable recorder was developed in the early 1970s. This technology was later adapted to EEG recording, and miniature preamplifiers that could be worn on the head were developed. [4, 5]

Early clinical investigations documented the ability of AEEG to record identifiable focal and generalized epileptiform activity. [6, 7] In 1982, Ives introduced a 16-channel AEEG that utilized signal multiplexing. [8] The 16 channels allowed improved spatial resolution and localization but recorded discrete samples rather than continuous EEG.

In 1983, a cassette tape AEEG system was introduced; it used off-head preamplifiers that had continuous 8-channel recording capability, real-time identification, and gain and filter adjustments. [9, 10, 11, 12]

In the past decade, computer technology has enabled portable recording of up to 36 channels with sampling rates of up to 400 Hz. Currently, numerous AEEG systems are available commercially (see the image below).

Components of an ambulatory EEG system. Components of an ambulatory EEG system.


AEEG has several important clinical applications. [13, 14, 15, 16, 17] Depending on the clinical suspicion, other diagnostic tests (eg, ambulatory cardiac monitoring, polysomnography [PSG] or inpatient video/EEG monitoring [18, 19] ) may be more appropriate in a given situation.

Confirmation of clinical suspicion of epilepsy

A clinical suspicion of epilepsy can be confirmed by recording a seizure on AEEG. [20, 21] This is most likely to occur when the patient is experiencing daily or almost daily spells. Studies looking at the diagnostic yield of AEEG indicate that 6-15% of AEEGs record seizures.

Higher yields have been reported from 16-channel AEEG with computer-assisted seizure detection than from older 4- or 8-channel systems without seizure-detection algorithms. [22, 23, 24, 25] A 2001 study in which 502 patients were evaluated with computer-assisted 16-channel AEEG demonstrated that 8.5% of patients had a seizure during the recording period (mean, 28.5 h). [26]

In patients with intractable epilepsy, AEEG has been used to localize seizure onset as part of presurgical evaluation. [27] However, inpatient video/EEG monitoring remains the standard for presurgical evaluation.

Evaluation of interictal epileptiform activity

Detection of interictal epileptiform abnormalities in the absence of recorded seizures can provide supporting evidence for a clinical diagnosis of epilepsy.

Studies have demonstrated that 34.9% of patients with known seizures had a positive AEEG, whereas 15.3% of 216 patients in whom the diagnosis of seizures was considered (ie, patients with episodic alterations of behavior, perception, sensation, or motor functioning) had interictal epileptiform abnormalities on 4-channel AEEG. When a 16-channel recorder was used, 38% of patients who were referred for AEEG had some type of epileptiform abnormality.

AEEG is highly specific; spikes were found on overnight AEEG in only 0.7% of asymptomatic adults without a history of migraine or a family history of epilepsy. [28] In patients with a history of migraine headaches and those with a family history of epilepsy, the incidence of spikes on AEEG was 12.5% and 13.3%, respectively.

Some patients in whom epilepsy is suspected have a normal routine or sleep-deprived EEG. In these patients, AEEG can increase the chance of detecting an epileptiform abnormality. Of patients who previously had normal or nondiagnostic routine EEG, 12-25% have epileptiform activity on AEEG.

A study comparing the usefulness of sleep-deprived EEG and computer-assisted 16-channel AEEG in patients with suspected epilepsy (but a nondiagnostic initial routine EEG) found that sleep-deprived EEG improved detection of epileptiform discharges by 24%, whereas AEEG improved detection by 33%. [29] Of the 46 patients studied, 15% had actual seizures recorded on AEEG, and none had seizures during the sleep-deprived recording.

Patients may have epilepsy without interictal epileptiform abnormalities on EEG, but this occurs in fewer than 20% of patients. In a study using a 4-channel recording system, 3 patients had only seizures recorded without interictal abnormalities. AEEG with 16 or more channels increases the probability that interictal epileptiform abnormalities will be found.

Documentation of seizures of which patients are unaware

For a patient to have seizures and yet be unaware of them is not uncommon. [30, 31] Brief alterations of awareness occur in both absence and complex partial seizures. AEEG is helpful at identifying seizures that are unrecognized or unreported by the patient.

Absence seizures may be so brief that the patient is unaware of them. A study using AEEG to evaluate absence seizures in pediatric patients found that most paroxysms of generalized spike and wave discharges (see the image below) were asymptomatic. [32]

An 8-second burst of generalized 3-Hz spike and wa An 8-second burst of generalized 3-Hz spike and wave captured on an ambulatory EEG.

Patients with complex partial epilepsy are often amnestic for their seizures. The sequelae of a nocturnal generalized convulsive seizure, if present at all, may be so subtle (eg, fatigue, muscle soreness) that the patient is unsure whether a seizure actually occurred.

A study of patients in an epilepsy monitoring unit found that 63% of all seizures were unrecognized by the patients. This difficulty in identifying the occurrence of seizures impedes seizure diagnosis and assessment of treatment adequacy.

Liporace et al found that the AEEGs of 3 patients (of 46) demonstrated seizures that were not designated as events by the patients. [29] Tatum et al found that more than one third of AEEGs with ictal activity contained at least one seizure that was unreported by the patient. [26] These studies demonstrate the utility of AEEG at capturing unsuspected events.

Evaluation of response to therapy

Because a significant number of patients are unaware of their seizures, responses to treatment are frequently difficult to gauge. Patients with mental retardation or other forms of encephalopathy may be unable to report seizures accurately. In such cases, AEEG can have a significant impact on clinical management.

AEEG is particularly useful in quantitating response to the treatment of absence seizures. If untreated, such seizures typically occur numerous times per day; adequate treatment usually normalizes the EEG.

Evaluation of nocturnal or sleep-related events

Certain diagnoses are difficult to confirm with the typical 20-minute outpatient EEG. The interictal epileptiform discharges of benign rolandic epilepsy, for example, are highly activated by sleep and may not always be achieved adequately in a laboratory. Continuous spike and wave activity during slow-wave sleep is another entity that may demonstrate a relatively normal EEG during waking hours and a strikingly abnormal EEG during deep sleep.

Because of its capacity to record an entire night of sleep, AEEG is invaluable in assessing these clinical situations. Another advantage is that children can be monitored at home. [33]

If a nonepileptic sleep disorder is suspected, PSG is the preferred study because of the added information from monitoring electromyography (EMG), eye movements, electrocardiography (ECG), and respiration.

The history may not differentiate clearly between a sleep disorder and epilepsy. AEEG may record frequent arousals (suggesting sleep apnea) or decreased rapid eye movement (REM) sleep latency (suggesting narcolepsy). In a study of 500 patients who had AEEG, narcolepsy was suggested in 6 patients, including 3 in whom narcolepsy had not been suspected.

Evaluation of suspected pseudoseizures

Pseudoseizures, also known as psychogenic seizures or nonepileptic events, [34] are clinical events in which patients perceive altered movement, emotion, sensation, or experience similar to those due to epilepsy but without an electrographic ictal correlate. [35, 36, 37, 38]

Pseudoseizures are surprisingly frequent, occurring in as many as 20% of patients at epilepsy referral centers and in 5-20% of outpatient populations. Some patients have both pseudoseizures and epileptic seizures; coincident events occur in an estimated 10-60% of epilepsy patients.

AEEG can be a useful screening tool in identifying patients who have nonepileptic paroxysmal events. In one study, 36% of patients had event marker activations without associated electrographic changes.

Potential problems exist in using AEEG to definitively diagnose nonepileptic seizures. A 24-hour recording without associated video does not allow evaluation of clinical stereotypy, which is valuable when evaluating patients with unusual seizure manifestations and minimal EEG changes. Scalp EEG may not show electrographic ictal abnormality during some frontal lobe seizures or may show only subtle abnormalities that would be difficult to interpret without associated video.

Kanner et al found that 25% of their group of 12 patients with supplementary motor seizures demonstrated no electrographic ictal pattern during seizures. [39]

Seizures and nonepileptic seizures may be associated with movement and muscle artifact that may obscure the underlying EEG. Although AEEG may be a useful initial screening tool for nonepileptic events, inpatient video/EEG monitoring remains the criterion standard in evaluating nonepileptic seizures.

Evaluation of syncope

AEEG may be helpful in evaluating syncope or near-syncope if an ECG lead replaces 1 of the EEG channels. [40] If cardiogenic syncope is suspected, a Holter monitor or prolonged cardiac event monitor may be more useful clinically. Although arrhythmias have been diagnosed with continuous ambulatory EEG/ECG recording, a study of epileptiform abnormalities in AEEG found that only 1 of 67 patients with syncope, near-syncope, or episodic dizziness had an epileptiform abnormality.

Future applications

Seizure anticipation methods are under development to identify EEG changes before seizure onset, allowing ongoing assessment of the probability of seizure occurrence. [41, 42, 43, 44, 35, 45, 46, 47, 48] With further characterization of EEG changes in the preictal state, future AEEG recording may be coupled with a seizure anticipation device, providing a time window within which therapeutic intervention may prevent a seizure.




Disk electrodes should be applied with collodion to ensure stability during the 24-hour recording period. Electrode gel may have to be reapplied by syringe beneath the electrodes. The patient’s scalp may be wrapped in gauze and the lead wires gathered and tacked to reduce traction on the electrodes (see the image below). New batteries should be provided every 24 hours.

Patient with scalp electrodes, carrying an ambulat Patient with scalp electrodes, carrying an ambulatory recorder.

Calibration is recorded using a 50-µV square wave or pulse; each electrode should be tapped sequentially. This part of the recording should be reviewed while the patient is in the laboratory to verify the integrity of the system and the appropriate connection of electrode leads to the preamplifier.

Patients should be instructed to record activity in a diary. Documenting physiologic artifacts, such as eye movements, chewing, speaking, and swallowing, is useful. Patients should not chew gum, which could produce a prolonged artifact. In addition, they should not bathe with the device.

When the patient returns to the laboratory for removal of the device, the end of the recording should be reviewed to ensure that signal recording was maintained.

In the past, when 4- and 8-channel AEEG was in use, montage design was of paramount importance in capturing and attempting to localize a suspected discharge. [49] Because abnormalities were recorded most often in the anterior temporal regions and frequently in the frontal regions, montages were designed to optimize coverage of these regions and to maximize yield.

With the newer digital 16- to 36-channel AEEG systems, standard EEG montages are usually adequate, since digital reformatting is always an option. Additional electrodes at T1/T2, zygomatic, and sphenoidal locations may allow more specific localization.

Review of Recordings

Because AEEG records a vast amount of data, scanning techniques are necessary to decrease review time. Cassette 8-channel systems allow scanning at speeds of 20, 40, and 60×. They also allow audio output monitoring. Although isolated, single discharges may be difficult to perceive at these speeds, seizures are less likely to be missed, because of the longer on-screen time, the ability to recognize of rhythmic or evolving patterns, and the presence of rhythmic audio bursts (usually of declining frequency).

Review of a 24-hour digital study would be prohibitively time-consuming without the benefit of computer-aided analysis. Newer digital AEEG models reduce reviewing time by means of sampling (which risks missing infrequent discharges) and automated spike and seizure detection programs. Pioneered by Gotman, these computer techniques identify epileptiform discharges on the basis of changes in amplitude, frequency, and rhythmicity. [50, 51, 52, 53]

Further refinements have reduced the incidence of false positives due to physiological artifacts and sleep patterns. Neural network modeling techniques may offer additional advantages in the future.

Because AEEG is done outside the controlled confines of the EEG laboratory and while the patient performs customary daily activities, it is susceptible to a variety of physiologic and environmental artifacts. Such artifacts may be difficult to recognize in AEEG, especially if a limited number of channels are in use.

Artifacts from sustained blinking, chewing, or movement may obscure the underlying EEG. Additionally, rhythmic artifacts due to repetitive activities (eg, teeth brushing, scratching, or bicycle pedaling) may mimic seizure activity. Ideally, the patient diary will document these occurrences. Judicious use of filters, gain, and reformatting may further clarify these waveforms and assist in distinguishing seizure activity from artifact.

Electroencephalographic seizures usually demonstrate evolution of amplitude and frequency, may spread to involve neighboring electrodes, and may be followed by postictal slowing or suppression. If doubt exists as to whether a discharge is artifactual or epileptic, a conservative interpretation should be followed.