Neonatal EEG

The neonatal EEG has some very different clinical considerations for recording and interpretation. Understanding certain clinical details, such as the conceptional (aka conceptual) age (CA) and the clinical state of the recorded patient, is essential for interpretation of the neonatal EEG.

The indications for the conventional neonatal EEG include assessment of age and maturity; identification of neonatal seizures and neonatal status epilepticus; evaluation of neonatal encephalopathy and focal abnormalities; and assessment of response to treatment or to aid neurologic prognosis. The conventional neonatal EEG is the gold standard for the diagnosis and confirmation of neonatal seizures and neonatal encephalopathy.

There are also several specific technical considerations for neonatal EEG, beginning with the montage and electrode placement. The neonatal montage is used from the time of birth until the baby reaches full-term age. In some centers, the neonatal montage is used until the baby is 46 to 48 weeks gestational age (GA) or until sleep spindles are seen in the recording (around 46–48 weeks) (see Figure 19).

Figure 19.

The 10-20 System electrode placements modified for neonates. Most of the neonatal EEG activity is found in the central regions of the brain, therefore the neonatal montage should have sufficient coverage of the centro-temporal regions. Figure courtesy (more...)

A study from Tekgul and colleagues compared the sensitivity and specificity of the reduced (neonatal) montage versus a full 10-20 montage in neonates (4). They found that the neonatal montage had a sensitivity of 96.8% and specificity of 100%. An electrode cap is used in some institutions in which there is no 24-hour EEG technologist coverage, since a cap can be placed by nurses, residents, or fellows. Electrocaps are color coded and can be adjusted to fit different head sizes. Other polygraphic parameters or extracerebral channels that are included in the conventional neonatal EEG are the electrooculogram (EOC), electromyogram (EMG), electrocardiogram (ECG), pneumograph, and video. For the EOC, two EOC electrodes are placed near the outer canthus of the eyes, one above the eye and the other below the eye. EOC allows for identification of different behavioral stages, in particular awake and active sleep stages, where eye movements are seen. For EMG recordings, the EMG electrode is placed under the chin. EMG allows for the identification of different behavioral stages (awake and active sleep), since active sleep is often associated with relative muscle atonia.

ECG leads are located on the chest to record variations of the heart rate and allow distinction of ECG artifact on the EEG. A pneumograph or respiratory belt also allows for the identification of behavioral stages. Synchronized video recording should also be used when possible, although a well-trained EEG technician or nurse annotating the EEG record can help substitute for tracking behaviors of the patient or environmental issues that may generate EEG artifacts, such as patting or nurse manipulation; this is crucial since sometimes movements such as these may generate artifacts that almost precisely mimic seizure patterns on the neonatal EEG.

Newborns, in particular preterm babies, have very thin and sensitive skin. Even when the recommendation is to keep the skin impedance (a measure of the quality of the connection between the skin and the recording electrode) at around 5 kΩ, an impedance of approximately 10 kΩ also may produce a technically adequate recording, while avoiding severe skin abrasions. The low-frequency filter is set lower in neonatal recordings than for EEG recordings in older children and adults to allow for the recording of slower frequencies at 0.005 to 0.01 Hz or 0.5 Hz, and the high-frequency filter setting is similar to adult recordings at 35 to 70 Hz.

Neonatal EEG recording should last at least 2 to 3 hours to capture awake and all sleep stages. Neonatal EEG is typically displayed with a longer time interval on the screen (a faster “paper speed” of 15 mm/s) producing a more compressed-appearing recording. This compressed screen allows for better display of very slow activity, asymmetries, and asynchronies that are crucial to evaluate in neonatal recordings.

Neonatal montages have some variations between institutions. The main variations are where the different channels are located on recording montages and how they are displayed on the screen or page (see Figure 20 for a typical neonatal montage display).

Figure 20.

Typical neonatal montage. Neonatal montages have some variations from institution to institution. The main variations are where the different channels are located. In this sample, the vertex electrodes are in the middle of the EEG trace, and the additional (more...)

To provide an accurate interpretation of the neonatal EEG, it is important to know the conceptional age (aka conceptual age) (CA) of the baby, the medications the baby is taking at the time of the recording, the different behavioral states of the baby, and any pertinent environmental changes. The CA is calculated by adding the estimated GA and the legal or chronologic age of the patient following birth. An example is a 4-week-old baby born at 30 weeks GA would have a CA of 34 weeks. Taking into account the CA, a neonate is a newborn infant with age <4 weeks. The definitions of preterm, near-term, and term are also important to become familiar with for neonatal EEG interpretation. A neonate is a newborn baby less than 4 weeks of age. A preterm baby has a CA between 24 and 34 weeks. Near-term babies have a CA between 34 and 36 weeks, while a term baby has a CA of 37 weeks and above.

Some medications and cooling therapy decrease the voltage of the neonatal EEG, so it is very important to know what medications and therapies are being given to the baby at the time of the EEG recording. Morphine, barbiturates, benzodiazepines, and other antiepileptic drugs decrease the voltage of the neonatal EEG. Head cooling and total body cooling also reduce the voltage of the neonatal EEG.

Technician annotations regarding the different behavioral states assist significantly in the interpretation of the neonatal EEG. Neonatal EEG recordings have clear differences during awake and sleep states, and within sleep stages. In general, neonates when awake have eyes open, whereas when they sleep, they have eyes closed. Regularity of respirations and eye movements help to differentiate between active sleep (REMs and irregular respirations) and quiet sleep (no REMs and regular respirations).

Sources of artifacts should also be noted by the technologist. Sources of artifacts in the neonatal EEG are ventilators, incubators, lines and drips, and feeding. Loud noises, flashes of light, and nursing or parental care can also be sources of artifacts and should be noted. These factors all produce transient attenuation of the neonatal EEG background as seen with arousals. Some artifacts can be produced in the EEG trace when EEG technicians are fixing the electrodes (Figure 21). These artifacts can be mistaken as sharp waves or even seizures. The same is true for patting artifact, that typically has a variable frequency from beginning to end and can resemble an ictal pattern (Figure 22).

Figure 21.

Neonatal EEG artifacts from technician fixing electrodes. This sample shows some artifacts (channels 3, 4, 7, 8, 11, and 12) produced by fixing the electrodes. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 22.

Patting artifact. This sample shows a widespread rhythmic artifact produced by patting the baby. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Similar to adult EEG, an orderly approach to visual analysis is necessary to effectively interpret neonatal EEG. The basic organization of the background rhythm should include an inspection of the EEG continuity and discontinuity, symmetry, synchrony, amplitude, reactivity, and morphology and composition of graphoelements.

Continuity in neonatal EEG refers to an EEG tracing with relatively constant and consistent amplitude (Figures 23, 24). Discontinuity in neonatal EEG refers to periods of relatively higher amplitude bursts that alternate with periods of lower amplitude, or interbursts (Figures 23, 25). The background evolves through different states in neonates with CA between 24 and 46 weeks, as shown in Table 2. Between CA 24 and 29 weeks, the EEG appears very similar in different states, and there is no reactivity to stimulation. The EEG is discontinuous but synchronous, and interburst intervals (IBI) are between 6 and 12 seconds with amplitude less than 2 μV. Between CA 30 and 34 weeks, the EEG has longer periods of continuity but is still relatively discontinuous and becomes somewhat reactive to stimulation. The EEG has a similar appearance during the awake and active sleep states. Quiet sleep is characterized by periods of discontinuity that are known as tracé discontinu. The EEG is synchronous in approximately 70 to 80 percent of the recording. From this age onward, the IBI intervals become progressively shorter, and the amplitude of the IBI progressively increases until the EEG becomes completely continuous around CA 44 weeks. Between CA 35 and 36 weeks, there is a clear distinction between the awake and active sleep states. The EEG is more continuous in both states (activité moyenne) but remains discontinuous during quiet sleep (known as tracé alternant, because of alternating periods of high-voltage burst intervals and low-amplitude IBI). The EEG is clearly reactive with voltage flattening, and increased continuity occurs during stimulation in quiet sleep. The EEG is more synchronous, during about 85% of the recording.

Figure 23.

Continuity of the neonatal EEG. Continuity in neonatal EEG refers to a trace with relatively steady amplitude. Discontinuity in neonatal EEG refers to periods of relatively higher amplitude or bursts that alternate with periods of lower amplitude or interbursts. (more...)

Figure 24.

Continuous neonatal EEG. Continuity in neonatal EEG refers to a trace with a steady amplitude. This sample shows a continuous EEG. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 25.

Discontinuous neonatal EEG. Discontinuity in neonatal EEG refers to periods of relatively higher amplitude or bursts that alternate with periods of lower amplitude or interbursts. This sample shows a normal 27-week CA infant with a discontinuous EEG, (more...)

Table 2.

Neonatal EEG Background Evolution in Different Behavioral States

Between 37 and 40 weeks CA, the EEG becomes continuous and appears similar during wake and active sleep states. During quiet sleep, there is tracé alternant (Figure 26) with some periods of continuous SWS. EEG is completely synchronous and reactive to internal or external stimuli. Between 40 and 44 weeks CA, the EEG is continuous during wake, active sleep, and continuous SWS portion of quiet sleep. EEG is reactive in all states and synchronous. Between 44 and 46 weeks CA, the EEG is continuous in all states. There is continuous SWS that replaces tracé alternant. Spindles appear in the central regions with a frequency of 12 to 14 Hz. Stimulation during continuous SWS produces relative attenuation of the EEG.

Figure 26.

Tracé alternant. This is a neonatal EEG sample of a 25-day-old girl born at 39 weeks GA. The sample shows a segment of quiet sleep with tracé alternant. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

These patterns of continuity and discontinuity are clinical-electrographically defined as tracé discontinu, around 30 to 34 or 35 weeks, with quiet periods of voltage <25 μV (often <10 μV); tracé alternant, occurring between 34 and 35 weeks until term during quiet sleep, with quiet periods of voltage >25 μV, alternating with bursts of 100 to 200 μV; and tracé continu, at 40 weeks CA and above, with continuous, irregular delta and theta of 50 to 100 μV during awake and active sleep.

Symmetry in neonatal EEG refers to symmetry of activity arising from both hemispheres or homologous brain regions. Elements to consider for evaluation of symmetry are amplitude, frequency, and waveform elements. Asymmetry is suspected when the amplitude of two homologous brain regions exceeds a ratio of 2:1 (Figure 27). When analyzing an asymmetric pattern, if asymmetry is only in amplitude, one should consider incorrect EEG placement, scalp edema, and subdural collections. If asymmetry is of frequency, amplitude, and graphoelements, then one should consider stroke or structural lesions.

Figure 27.

Asymmetric background. This neonatal EEG sample was recorded in a 2-day-old full-term baby boy who had a left stroke. Notice the asymmetry in amplitude and frequency seen in the electrodes covering the left hemisphere (oval). This was an intermittent (more...)

Synchrony has several meanings in EEG interpretation but in this case refers to the interhemispheric timing of graphoelements, mainly during the discontinuous portions of the neonatal EEG. EEG bursts are considered synchronized when there is less than 1.5 seconds separating the onset of the burst between the two hemispheres. Graphoelements that are always synchronous are encoches frontales and anterior frontal dysrhythmia, which are seen in all behavioral states but especially during quiet sleep in the transition from active to quiet sleep (Figure 28), and monorhythmic occipital delta. The amount of synchronization varies during neonatal EEG maturation. Before 29 to 30 weeks CA, bursts are 100% synchronous. Synchrony decreases to approximately 70% between 31 and 36 weeks CA and increases progressively thereafter until again reaching 100% at age 37 weeks CA. Bursts are asynchronous if more than 1.5 seconds separate the onset of the bursts between the right and left hemispheres (Figure 29). Asynchrony can be seen in any condition that causes diffuse encephalopathy, and in cerebral dysgenesis with callosal agenesis.

Figure 28.

Neonatal EEG synchrony: example of encoches frontales and anterior frontal dysrhythmia. This is a sample of quiet sleep in a 25-day-old baby girl born at 39 weeks GA. There are two synchronous graphoelements in this sample: anterior frontal dysrhythmia, (more...)

Figure 29.

Neonatal EEG: asynchrony. This is a sample of an abnormal neonatal EEG from a 2-week-old baby born at 37 weeks having neonatal seizures as a result of hypoxic-ischemic encephalopathy and sepsis. Notice the asynchrony of the bursts. The EEG is also significant (more...)

The amplitude of the EEG is measured in voltage. The voltage value is measured from peak to peak of the waveform. In neonatal EEG, the amplitude of the graphoelements decreases from 24 weeks CA to term. Amplitude abnormalities include an isoelectric EEG, a depressed or undifferentiated EEG with voltage less than 10 μV, or an EEG with persistent low voltage under 5 to 10 μV when awake, under 10 to 25μV during quiet sleep, or low voltage persistent beyond 43 weeks CA (Figure 30a and b).

Figure 30.

Low-amplitude neonatal EEG in hypoxic-ischemic encephalopathy. (a) Low-amplitude, sleep, newborn with hypoxia; (b) low-amplitude neonatal EEG: arousal, no change in stage. (a) and (b) correspond to an EEG sample of a full-term baby boy with severe hypoxic-ischemic (more...)

Reactivity is the clinical or EEG response to external stimulation or internal arousal. There are clinical changes and EEG changes that indicate reactivity. Clinical response includes active movements and respiratory pattern changes. EEG response includes frequency changes, increased continuity, decreased amplitude, and change from sleep to a wakeful pattern (Figure 31). Photic stimulation does not produce photic driving in the term neonate. Absence of reactivity is normal in premature babies under 30 weeks CA. Otherwise, absence of reactivity indicates pathological thalamo-cortical disruption.

Figure 31.

Reactivity during arousal. This sample corresponds to a neonatal EEG of a full-term baby. Notice the reactivity of the EEG to arousal with a relative voltage attenuation (seconds 9–14) followed by continuous activity (second 15 until end of page). (more...)

Table 3 shows the development of the different graphoelements at different CAs. When interpreting neonatal EEGs, it is important to learn to differentiate normal sharp-wave transients from sharp waves that can be indicative of neonatal CNS dysfunction. Sporadic sharp waves are present during all preterm and term recordings. Examples include encoches frontales (Figures 28, 32) and sharp transients located in the centro-temporal regions (Figures 33 and 34). Many sharp transients in neonatal EEGs can be considered artifacts until proven otherwise (Figure 35). As in adult EEGs, abnormal sharp waves are most often negative polarity sharp waves with a classic morphology, including a cerebral electrical field and an after-going slow wave that disrupts the background. The relation of neonatal sharp waves with neonatal seizures and subsequent risk for epilepsy is often unclear. Sharp waves seen in the occipital region and midline are usually abnormal. Positive sharp waves generally have no relation to seizures but are instead related to structural brain abnormalities; however, though rare, positive sharp waves may be epileptogenic (Figure 36). When positive sharp waves are located in the rolandic areas, they are most often associated with white matter lesions.

Table 3.

Development of Graphoelements

Figure 32.

Encoches frontales. This EEG sample shows a normal sharp transient, encoches frontales. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 33.

Normal sharp transients, central region. This EEG sample shows normal sharp transients in the central region. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 34.

Central delta brushes, right rhythmic temporal theta. This is an EEG sample of a 32-week CA baby during sleep. Notice the centrally located delta brushes (delta wave with superimposed alpha beta activity 8 to 20 Hz) and the right temporal theta (brief (more...)

Figure 35.

Neonatal EEG: artifact. Sharp wave–like transients owing to artifact in T7. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 36.

Positive temporal sharp wave. This EEG sample is from a 2-week-old baby, born at 37 weeks, who had neonatal seizures. The sample shows positive temporal sharp waves. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Sleep/wake cycles can only be differentiated by EEG after 31 weeks CA. Awake and active sleep have some clinical and EEG similarities, including irregular respirations and mixed frequencies in the EEG background. In contrast to older infants and adults, active sleep follows wakefulness in neonates. Active sleep accounts for up to half of the sleep time in neonates. Clinically, quiet sleep is characterized by absence of eye movements in the EOG, regular respiration, and absence of movement artifacts. On the EEG, trace alternant is characterized by quiet periods of voltage over 25 μV, alternating with bursts of 100- to 200-μV amplitude. Slow quiet sleep shows continuous high-amplitude delta activity over all brain regions. Encoches frontales is seen during quiet sleep. Transitional sleep or undetermined sleep cannot be classified into active or quiet sleep and is mainly seen between 37 and 40 weeks CA, during transitions between the different behavioral states.

In conclusion, analysis of the neonatal EEG background begins with knowledge of the conceptional (conceptual) age and clinical state of the recorded neonate. Subsequent interpretation should include an assessment of the EEG background continuity, symmetry, synchrony, normal and abnormal patterns, sleep/wake cycle, and seizures.

Infant and Pediatric Developmental Changes in the EEG

Thus far, we have considered the EEG in preterm and term neonates below 38 weeks GA. Infants may be defined as being in the age period between 1 and 12 months; toddlers, in the 1- to 3-year span; and preschool children, in the 3- to 6-year age range. School-age children are then in the 6- to 18-year span, with further changes in the EEG occurring in the subdivision of children aged 6 to 12 years and in teenagers aged 13 to 19 years.

During infancy (1–12 months), there are specific changes in the EEG background. By age 2 months, a posterior dominant rhythm (PDR), a forerunner of the alpha rhythm, is established. It usually begins as a 3- to 4-Hz frequency, increasing to 4 to 5 Hz by age 6 months, reaching approximately 5 to 7 Hz by 12 months (Figures 37–40), and finally becoming an alpha frequency range of 8 Hz by 3 years. Transition between wakefulness and drowsiness is apparent when the background slows by 1 to 2 Hz and fronto-central activity may predominate and reach relatively high amplitudes of approximately 200 μV. Sleep spindles typically develop by 2 to 3 months, are often asynchronous, and may remain so until about 6 to 12 months of age. During the age period of 6 to 12 months, sleep spindles can have very long durations, lasting for 10 to 15 seconds, and may be quite asynchronous by durations as long as 1 to 5 seconds (Figure 41). V-waves and K-complexes develop between 2 and 5 months, and a slower frequency background of 1 to 3 Hz predominates during sleep. At this age, V-waves are characteristically sharp and spiky, often occur in repetitive trains, and appear lateralized with a fronto-central or even central distribution, requiring caution to carefully distinguish them from epileptiform discharges. REM sleep begins to diminish from approximately 50% levels of active (REM) sleep seen in newborns, to about 40% of sleep time by age 3 to 5 months, reaching 30% by 12 to 24 months of age. Activating procedures, such as hyperventilation, are not practical until about 3 years when children will cooperate with instructions. However, during spontaneous crying, state transitions between sleep and arousal to wakefulness, or during periods of drowsiness, a prominent buildup of slower high-voltage theta and delta EEG frequencies may be seen, similar to the effect of hyperventilation seen in children and adolescents (Figure 42). Photic stimulation may begin to produce a slow-frequency driving response of 1 to 3 Hz during infancy by about 6 months, and photic responses may be exaggerated in certain encephalopathies including neuronal ceroid lipofuscinosis and Gaucher disease.

Figure 37.

Pediatric EEG: 5- to 6-Hz PDR in an 11-month-old child. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 40.

Pediatric EEG: eye closing in a 23-month-old boy playing “peek-a-boo,” demonstrating 7-Hz PDR. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 41.

Pediatric EEG: asynchronous spindles in a 16-month-old boy. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 42.

Frontal arousal in a 16-month-old boy. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 38.

Pediatric EEG: 6- to 7-Hz PDR in a 12-month-old child. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 39.

Pediatric EEG: normal 16-month-old waking EEG record, reaching 7- to 8-Hz PDR. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Many further maturational changes unfold during the first 3 years of life. The PDR reaches about 8 Hz by 3 years in most toddlers, but the predominant EEG background frequencies remain slower in the delta and theta frequency ranges of 2 to 6 Hz. A prominent μu rhythm over the central regions usually develops by 1 to 2 years, typically before the alpha rhythm fully reaches 8 Hz, and beta frequency activity also emerges. During drowsiness, hypnagogic hypersynchrony, a buildup of high-voltage delta activity, which can have sharp or spiky components often in the fronto-centro-parietal regions, is common following the first year of life and especially around 2 to 4 years. This should not be mistaken for epileptiform activity. By 1 to 3 years of age, sleep spindles should be synchronous and symmetric, and reach a frequency of 12 to 14 Hz.

In preschool children aged 3 to 6 years, theta frequencies remain in the background, but the alpha PDR background frequency increases further until reaching 8 to 9 Hz by 5 to 8 years of age. This is often intermixed with delta activity known as posterior slow waves of youth, which may persist into teenage years and even into young adulthood. Further maturation of the alpha rhythm frequency may still occur, with alpha activity reaching 8 to 9 Hz by 8 years of age, 9 to 10 Hz by 10 years, and adult range frequencies of 8 to 12 Hz by age 12 to 13 years (Figure 43). By 16 years of age, the minimal background alpha frequency should be 8.5 Hz, although posterior slow waves of youth may persist into the late 20s. Common NREM sleep features including V-waves, K-complexes, and sleep spindles become fully developed in school-aged children (Figure 44). Hyperventilation may now be routinely performed to attempt to activate epileptiform activity and absence seizures, and frequently elicits a prominent slow wave build-up response (Figures 45, 46). Photic stimulation may routinely elicit a posteriorly predominant driving response (Figure 47), although absence of driving is not abnormal at any age. Lambda waveforms elicited by complex pattern viewing also appear in preschool-aged children between 3 and 5 years of age and become more frequent from 6 to 12 years of age. At the same time, POSTS appear during sleep, and more prominent photic driving responses at faster flash frequencies emerge. Anterior rhythmic theta bursts of 6- to 7-Hz frequency are often seen, especially during drowsiness in the mid-teenage years between 13 and 16, as prominent hypnagogic hypersynchrony begins to diminish.

Figure 43.

Pediatric EEG: normal waking background in a 10-year-old girl, showing a 10-Hz alpha PDR. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 44.

Pediatric EEG: N2 (stage 2 NREM) sleep in an 8-year-old boy. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 45.

Pediatric EEG: buildup of slow wave frequencies during hyperventilation in a 7-year-old girl. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 46.

Pediatric EEG: build-up response becomes more prominent later during hyperventilation in the same 7-year-old girl as shown in Figure 43. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.

Figure 47.

Pediatric EEG: photic driving response at 10-Hz flash frequency in a 7-year-old girl. Figure courtesy of Elia M. Pestana-Knight, MD, Cleveland Clinic Foundation.