Neonatal EEG

2196

Neonatal EEG

Natrujee Wiwattanadittakun and Tammy N. Tsuchida

LEARNING OBJECTIVES

• Develop a systematic approach to the assessment of the neonatal EEG

• Understand the maturation of the EEG from preterm to term

• Identify normal patterns and graphoelements for gestational age

• Differentiate burst suppression from an excessively discontinuous tracing

• Recognize that the effects of therapeutic hypothermia and medication may impact the neonatal EEG background

Introduction

Conventional electroencephalography (EEG) is the gold standard for detecting seizures and for assessing the EEG background in neonates. Neonatal EEG interpretation is age-dependent, with background patterns changing markedly from preterm to term gestation. This chapter will cover the approach to the visual analysis of normal and abnormal neonatal EEG background patterns.

Neonatal EEG Recording Techniques

Electrode Placement

Neonatal electrode placement is adapted from the international 10-20 system.¹ While the same number of electrodes used for an older child can also be used for a neonate, it is common to use a reduced set of electrodes in a “double distance” montage.¹ The American Clinical Neurophysiology Society (ACNS) guidelines allow for more posterior placement of the frontal electrodes—AF3 (halfway between Fp1 and F3) and AF4 (halfway between Fp2 and F4).^1,2 Other electrode placements are similar to those used in older children and are shown in Figure 6.1.¹ Gold and silver electrodes are most commonly used. Subdermal wire electrodes may enable longer term recordings but are much less commonly used. Dry sensor electrodes, which may reduce skin injury, are also under investigation.³

Noncerebral Channels¹

Noncerebral channels record eye movements, respiration, cardiac activity (electrocardiogram [ECG]), and muscle activity. The monitoring of eye movements occurs via electrodes placed obliquely across either eye to detect vertical and lateral eye movements, the monitoring of muscle activity occurs via electrodes placed under the chin, and the monitoring of respiratory function occurs via a monitor placed on the abdomen or chest. Respiratory, eye movement, and muscle activity channels are important for the accurate assessment of behavioral state. The addition of noncerebral channels can also aid in deciphering artifact from abnormal cerebral activity. At a minimum, ECG and respiratory leads should be included in any neonatal montage. Figure 6.2A indicates the order of cerebral electrodes in a typical neonatal double distance montage, while Figure 6.2B includes the noncerebral channels.

**FIGURE 6.1.** **Double distance montage.** Schematic illustrating the reduced set of electrodes often used in the neonatal montage. **(A)** The left image shows the electrode placement using International 10-20 system nomenclature. The green electrodes represent a possible alternative placement for the frontal electrodes and the yellow electrodes highlight the electrodes used in a limited neonatal montage. The black arrows indicate the ordering of electrodes in a typical neonatal montage as a combination of longitudinal bipolar chains (Fp1 → T3 → O1, Fp1 → C3 → O1, Fp2 → C4 → O2, and Fp2 → T4 → 02) and a transverse bipolar chain (T4 → C4 → Cz → C3 → T3). **(B)** The right image shows the same electrode placement using International 10–10 system nomenclature.

220Video Recording

Time-locked video recording is an essential component of the neonatal EEG. It allows for the assessment of subtle clinical correlates to electrographic seizures, aids in determination of behavioral state, and assists in differentiating cerebral activity from artifact. This is especially helpful in the critically ill neonate undergoing frequent nursing cares and interventions, where certain movement artifacts, for example, patting, may mimic a seizure on the EEG (Figure 6.3). (More details on neonatal recording artifacts are discussed in Chapter 7.)

Technical Aspects

Electrode impedances for neonates can be slightly higher than those recommended for older children to help minimize skin breakdown.¹ Nonetheless, impedances should remain under 10 kOhms, and marked differences in impedance between electrodes should be avoided.¹ Paper speed is typically set at 15 mm/s, instead of the standard 30 mm/s used for older children and adults, as this can facilitate assessment of the degree of discontinuity as well as the detection of slowly evolving seizures (Figure 6.4). Sensitivity is typically set at 7 µV/mm but may need to be changed to 10 µV/mm or 15 µV/mm to enable effective interpretation of high amplitude tracings.¹ In contrast, a sensitivity of 2, 3, or 5 µV/mm may be needed to evaluate a low voltage background, such as may be seen after an acute insult. The low-frequency filter should be set at 0.3 to 0.6 Hz (time constant 0.27–0.53 seconds) to capture normal neonatal delta activity.¹ The high-frequency filter is set at 70 Hz, as for older children.¹

Collaboration between the EEG technologist and bedside provider is critical to obtaining a successful neonatal EEG recording.¹ EEG technologists and bedside providers can help to minimize electrical and other artifacts. They can also make notation of behavioral states as well as clinical characteristics and events, such cyanosis and apnea, that are not easily visualized on video, especially in low ambient lighting.¹ Ideally, the technologist should stimulate the newborn and note this in the recording to enable the EEG reader to determine EEG reactivity. This should include auditory and tactile stimulation. Photic stimulation is not mandatory in neonatal EEG recordings.¹

EEG Duration

The recommended duration of the EEG depends on the objectives of the recording. ACNS recommends at least 60 minutes for background assessment as this is adequate to capture a normal sleep wake cycle.¹ If the goal is to capture seizures in a high-risk newborn, at least 24 hours is recommended.^4,5 A shorter duration may be adequate for a lower-risk newborn with a normal or mildly abnormal background.⁶

Visual Analysis of the Neonatal EEG

Important Clinical Information

Important information that should be obtained prior to visual analysis of the neonatal EEG includes postmenstrual age (PMA), behavioral state, and whether the neonate is being treated with neuroactive medications and/or therapeutic hypothermia.^1,7 The American Academy of Pediatrics (AAP) has endorsed the concept of postmenstrual age (PMA), rather than conceptual age (CA);^1,7 however, CA is widely used in older literature. PMA is the summation of gestational age (GA) and chronological age, where GA is measured from the time of the last menstrual period.^1,7 Since the last menstrual period is approximately 2 weeks prior to conception, PMA roughly correlates to CA plus 2 weeks.^1,7 Term neonates are defined as born at 37 weeks PMA or after but before 44 weeks PMA.⁷ Preterm neonates are born before 37 weeks PMA, and postterm neonates are born at 44 weeks PMA or after.⁷ In the case of an unreliable menstrual history or prenatal ultrasound, the PMA as determined from the EEG background might be more accurate than the PMA determined from clinical history alone. EEG background features can typically be used to estimate the PMA to within 2 weeks.

Overview of Visual Analysis

A systematic approach to the assessment of the neonatal EEG background is essential to effective interpretation of the EEG. First, as noted above, it is important to know the postmenstrual age of the neonate as normal background features change based on PMA. Second, it is important to be able to recognize a continuous versus a discontinuous pattern (Figure 6.5). While this may appear relatively simple, recognition of these two patterns is the first step in determining if the electrographic state of the neonate matches the clinical or behavioral state. The remaining elements of the neonatal EEG that should be assessed in a systematic fashion include state change, voltage, symmetry, synchrony, reactivity, variability, the presence or absence of normal graphoelements, and the frequency of sharp transients.⁷ Each component is covered in greater detail below.

Continuity

The EEG background matures with age, progressing from discontinuous in all states, to continuous in active sleep, then wakefulness, and finally quiet sleep.⁷ There are variable definitions of discontinuity. The ACNS defines discontinuity as activity interrupted by an interburst interval (IBI) with background attenuation <25 to 50 µV peak to peak (pp) amplitude for 2 seconds or more.⁷ Another definition of discontinuity is EEG activity interrupted by an IBI with activity <25 µV pp for more than 3 seconds.⁸ EEG sharp transients in multiple electrodes or activity in a single electrode above 25 µV pp can be seen during the IBI.⁷ The normal discontinuous patterns in the preterm and term neonate are called tracé discontinú (Figure 6.6A) and tracé alternant, respectively. Continuity is defined as uninterrupted electrical activity with less than 2 seconds of voltage attenuation <25 µV pp (Figure 6.6B).⁷

Electrical activity first arises in fetuses by the eighth week of gestation.⁹ The ontogeny of EEG activity begins with brief periods of cerebral activity (bursts) alternating with periods of minimal to no activity (interburst intervals). The duration of the IBI decreases with age; IBIs are up to 46 seconds at a PMA of 26 weeks and decrease to 2 seconds near term.^8,10,11 Normal discontinuity may be seen in quiet sleep past 42 weeks, with fragments of discontinuity seen up to 46 weeks PMA.⁷ Table 6.1 lists ACNS-proposed maximal IBI durations and amplitudes for PMA.⁷ The burst and interburst interval (IBI) also vary with the behavioral state of the neonate and can be altered by brain injury and medications (see below). Importantly, if there is only a single or a few periods of discontinuity, it should be determined whether or not this is due to reactivity; neonates can have attenuation of activity <25 µV pp with spontaneous arousal or in response to external stimulation (see below) (Figure 6.7).

TABLE 6.1 NORMAL IBI DURATION AND AMPLITUDE

PMA (weeks)	Maximum IBI (seconds)	Voltage of IBI (µV)
<30	35	<25
30–33	20	<25
34–36	10	~25
37–40	6	>25

Source: Reproduced with permission from Tsuchida TN, Wusthoff CJ, Shellhaas RA, et al. American Clinical Neurophysiology Society standardized EEG terminology and categorization for the description of continuous EEG monitoring in neonates: Report of the American Clinical Neurophysiology Society Critical Care Monitoring Committee. J Clin Neurophysiol. 2013;30:161–173.

Behavioral and Electrographic State Change^7,8,12

OVERVIEW

The three clinical states of a term neonate—awake, active sleep, and quiet sleep—can be distinguished by both behavioral characteristics and electrographic patterns. This same principle holds for all age groups. For example, the awake state in a child is marked by muscle activity and a posterior dominant rhythm, while sleep is associated with loss of muscle activity, vertex waves, and sleep spindles. There is a greater emphasis on determination of these states in newborns, and sleep wake cycling is a positive prognostic sign in critically ill neonates. As detailed above, determination of the behavioral state relies on observation of the neonate via time-locked video recording as well as on the noncerebral channels used to detect eye movements, muscle tone (chin leads), and respirations (regular or irregular). We will review each clinical state, focusing on the behavioral and EEG characteristics (Figure 6.8).

STATE CHANGE IN TERM NEONATES (>37–<42 WEEKS)

Awake

Behavioral characteristics. In the awake state, the eyes are open and respirations are irregular. Movement artifact is present and associated with sucking and crying. The awake state can be further differentiated into a quiet awake state and an active awake state. EEG characteristics. The EEG background of the awake term neonate is continuous, low to medium in amplitude (25–50 µV), and comprised of theta and delta activity with overriding beta activity.⁷ This pattern is called activité moyenne (“average activity”) and is first seen in the awake state at approximately 34 weeks (Figure 6.9).

Active Sleep

Behavioral characteristics. In active sleep, the eyes are closed with periods of rapid eye movement, respirations are irregular, and there are intermittent body movements. Electrographic characteristics. There are two distinct patterns of active sleep. The first is a mixed frequency pattern (activité moyenne) and the second is a low-voltage irregular pattern (LVI). In the term neonate, active sleep is typically preceded by the awake state, and activité moyenne typically precedes the LVI pattern during a normal sleep-wake cycle. Active sleep ultimately becomes REM sleep (Figure 6.10).

Quiet Sleep

Behavioral characteristics. In quiet sleep, the eyes are closed, rapid eye movements are not seen, respirations are regular, and there is a paucity of movement.

Electrographic characteristics. There are two distinct patterns of quiet sleep in the term neonate. (a) Tracé alternant is the predominant pattern of quiet sleep between 37 and 40 weeks PMA and consists of bursts of 50 to 150 µV pp delta activity admixed with higher frequency activity superimposed on 25 to 50 µV pp theta and delta activity. This is a normal discontinuous pattern in the term neonate, becomes much less prominent by 42 weeks PMA, and is no longer present after 46 weeks PMA. (b) Continuous slow wave sleep/Slow continuous tracing. As tracé alternant recedes with increasing gestational age, quiet sleep has increasing amounts of continuous 50 to 150 µV pp delta and theta activity, the mature quiet sleep pattern that has been called continuous slow wave sleep or slow continuous tracing. This is the only quiet sleep pattern seen by 45 to 46 weeks PMA.¹³ Quiet sleep ultimately becomes slow wave sleep or N3 sleep (Figures 6.11–6.13).

Transitional Sleep

Transitional sleep is the period of sleep between states. It may have both behavioral and electrographic characteristics of the preceding and ensuing states (Figure 6.14).

Indeterminate Sleep

Indeterminate sleep is a sleep state during which the eyes are closed but the behavioral and electrographic characteristics prohibit a distinction between active or quiet sleep. In healthy term neonates, one quarter of sleep is indeterminate.¹⁴ In critically ill neonates, the indeterminate state is more commonly seen and may be the predominant sleep state.¹⁴

Sleep-Wake Cycling¹⁵

A typical neonate has 50 to 60 minute (range 30–70 minute) sleep cycles that alternate with waking episodes in a 3-to-4-hour cycle throughout the day. The first phase in a sleep cycle is active sleep, then quiet sleep, then active sleep, with transitional sleep between each phase. Each sleep cycle consists of about 50% to 60% active sleep, 30% to 40% quiet sleep, and 10% to 15% transitional sleep. Neonates under 35 weeks PMA have slightly shorter sleep cycles with a greater proportion of indeterminate sleep (Figure 6.15).

Unspecified State Changes

If a neonate has state cycling consisting of distinctly different EEG patterns (i.e., continuity, voltage, and frequency) that are not identifiable as active or quiet sleep states, ACNS neonatal terminology has defined this as unspecified state changes. At least 1 minute must be spent in each unspecified state (Figure 6.16).

**FIGURE 6.15.** **Sleep-wake cycling.** Established sleep-wake cycling at term follows a distinct pattern starting with the awake state (A), then active sleep, quiet sleep, and then active sleep with transitional sleep (TS) states in between.

Source: Image courtesy of Arnold Sansevere.

STATE CHANGE IN PRETERM NEONATES

EEG and polysomnographic studies have been performed on neonates as young as 24 weeks PMA. Prior to 24 weeks, behavioral states cannot be distinguished as the eyelids remain fused.

24 to 29 weeks PMA. The EEG background is discontinuous despite the behavioral appearance of distinct sleep and wake states (Figure 6.17).

30 to 33 weeks PMA. Active sleep becomes continuous while the awake state and quiet sleep remain discontinuous (Figure 6.18).

34 to 36 weeks PMA. The background during the awake state becomes continuous, while quiet sleep remains discontinuous (tracé discontinú) (Figure 6.19).

Tracé discontinú is the electrographic correlate of quiet sleep in the preterm neonate at approximately 34 to 36 weeks PMA, once the awake state becomes continuous. This pattern consists of bursts of high amplitude (50–300 µV pp) activity alternating with interburst intervals <25 µV, the duration of which varies by PMA. The tracé discontinú pattern is also seen prior to the awake state and active sleep becoming continuous. Tracé discontinú begins to transition to tracé alternant at approximately 35 to 36 weeks, as the IBI exceeds 25 μV.

Voltage^7,8,12

In the preterm neonate, attempts to classify normal voltage are limited by the lack of published norms. In the healthy term neonate, EEG activity should be >25 µV pp in all behavioral states. A pattern with some normal EEG features for age and a voltage of at least 10 µV but less than 25 µV is defined as borderline low voltage and has unknown clinical significance. A tracing without normal features and a voltage of <10 µV pp is called low voltage suppressed or low voltage undifferentiated. This is an invariant, unreactive pattern with baseline voltage <10 µV pp, which can be interspersed with voltage ≥10 µV pp for <2 seconds. This is in contrast to electrocerebral inactivity (ECI) in which there is no cerebral activity ≥2 µV.

Symmetry^7,8,12

The EEG background should be symmetric at all gestational ages. Voltage, frequency, graphoelements, and waveform complexity should be similar in homologous regions. A persistent difference in voltage exceeding a 2:1 ratio is abnormal.^7,16–18 Brief, transient periods of asymmetry are considered normal. False asymmetries may be present in the setting of extra-axial fluid collections, cephalohematomas, and/or improper lead placement (Figure 6.20).

Synchrony^7,8,12

As with the degree of continuity, the degree of interhemispheric asynchrony changes with brain maturation. Interhemispheric asynchrony is measured during the discontinuous portion of the EEG. Asynchrony is considered present when the onset of the bursts of cerebral activity in the two hemispheres occur more than 1.5 seconds apart. Neonates less than 27 to 29 weeks PMA have synchronous activity. The degree of synchrony initially decreases as the neonate matures and nadirs at 70% synchrony at 29 to 30 weeks PMA. The degree of synchrony then increases and is nearly 100% by 37 weeks PMA. Note that some asynchrony may be present in quiet sleep in the term neonate (Figures 6.19A and C).

Reactivity^7,8,12

Reactivity is defined as a change in the EEG background in response to an external stimulus. This change may occur in continuity, frequency, and/or voltage. Reactivity can be seen as early as 30 weeks PMA, is reliably obtained by 35 to 36 weeks PMA, and demonstrates consistency in the type of EEG change seen by 37 weeks PMA. The absence of reactivity is vital to the proper classification of patterns such as burst suppression and abnormally low voltage tracings.

Auditory and tactile stimulation should be performed at the start of the recording and repeated as clinically warranted to assess for reactivity. The electroencephalographer should be cautious when evaluating reactivity, as a behavioral response to stimulation can produce muscle artifact that may be mistaken for a cerebral response; muscle activity without any other EEG change does not qualify as reactivity (Figure 6.21).

Variability^7,8,12

Variability (lability) refers to an EEG background change in response to an internal stimulus. As with reactivity, this change may occur in continuity, frequency, and/or voltage. For example, a spontaneous arousal from sleep 224can result in a brief period of background attenuation. Variability can be seen earlier than reactivity, at 25 weeks PMA, and is established by 30 to 31 weeks PMA.

Normal Graphoelements^7,8,12

Graphoelements are brief EEG patterns and, in some cases, individual waveforms that are seen at specific postmenstrual ages. They are a marker of brain health. When they are seen at the expected gestational age, they are normal. As with other elements of the neonatal EEG, they should be symmetric. When normal graphoelements are seen outside of their expected gestational age range, they may be a marker of dysmaturity (Figure 6.22).

MONORHYTHMIC DELTA ACTIVITY

This pattern consists of runs of 0.5 to 1 Hz monomorphic, high-amplitude (up to 200 µV pp), surface-positive delta waves. These are seen in neonates as early as 23 to 24 weeks PMA and increase in prominence until 31 to 33 weeks PMA, before disappearing around 34 to 35 weeks PMA. Monorhythmic delta activity lasts for 2 to 60 seconds and is symmetric and synchronous. The delta activity is seen primarily in the occipital areas, at times with spread to the temporal and central, but not frontal, regions (Figures 6.23A–C).

RHYTHMIC OCCIPITAL THETA ACTIVITY

Runs of sinusoidal 4 Hz theta activity in the occipital area can be seen admixed with monorhythmic delta activity. The runs are shorter, lasting for approximately 2 to 10 seconds, and at times, spread to the temporal area. Similar to monorhythmic delta activity, this can be seen in neonates as young as 23 to 24 weeks PMA before peaking at 30 weeks PMA and then disappearing around 33 weeks PMA (Figures 6.23A–C).

SHARP THETA ON THE OCCIPITAL AREA OF PREMATURES¹⁹

The highest incidence of sharp theta on the occipital area of prematures (STOP) is at 23 weeks PMA. The incidence then declines between 27 and 30 weeks, only to peak again between 30 to 35 weeks PMA before slowly declining by term. This pattern has characteristics similar to temporal theta bursts of the premature. The two patterns can be distinguished based on their location, age of onset, and the unilateral nature of STOP (temporal theta of the premature is typically bilateral). STOP has been associated with seizures and with intraventricular hemorrhage, especially in the presence of positive sharp waves. It may be difficult to decipher STOP from rhythmic theta in the occiput, but it seems that these are distinct entities. The primary difference is the sharpness of the theta in STOP (Figures 6.23D–E).

DELTA BRUSHES

This pattern can be seen between 26 and 38 weeks PMA with peak prominence at 32 to 34 weeks PMA. Delta brushes are comprised of a slow wave (0.3–1.5 Hz) with superimposed fast activity (8–12 Hz and 18–22 Hz with the faster frequency range being most prominent prior to 32 weeks PMA). Delta brushes appear asynchronously²⁰ and show a variable voltage asymmetry. They appear initially during both awake and asleep states. Before 33 to 34 weeks PMA, delta brushes are most prevalent in active sleep; after this age, they are more prevalent in quiet sleep. By early term, delta brushes are no longer seen in wakefulness, and by 40 weeks PMA, they occur only during quiet sleep. They are rare by 42 weeks PMA and vanish by 44 weeks PMA. Delta brushes localize to the central regions until approximately 30 weeks, at which point they migrate to the temporal and occipital areas. Of note, delta brushes are not the precursor of sleep spindles, which first appear in quiet sleep at approximately 46 weeks PMA. Other names include “spindle-delta bursts,” “brushes,” “spindle-like fast waves,” “ripples of prematurity,” “beta-delta complexes,” and “rapid bursts” (Figure 6.24).

TEMPORAL THETA

This pattern has the same features as occipital theta activity but is sharper in contour and more rhythmic. The sharp contour and rhythmic nature of temporal theta can easily be mistaken for an ictal discharge. However, temporal theta does not evolve and appears bilaterally, albeit independently. Temporal theta can be seen between 24 to 34 weeks PMA with a peak at 30 to 32 weeks.²⁰ Temporal theta is replaced by temporal alpha, a specific marker for 33 weeks PMA that is no longer present at 34 weeks PMA (Figure 6.25).

ANTERIOR DYSRHYTHMIA

Anterior dysrhythmia consists of brief bursts of medium amplitude, 50 to 100 µV pp, synchronous, semirhythmic delta activity that is most prominent in the bifrontal areas. Anterior dysrhythmia is seen in neonates between 32 and 44 weeks PMA. While it can appear in all behavioral states, it is most abundant in the transition from active sleep to quiet sleep. Anterior dysrhythmia can be seen with admixed encoches frontales (see below). Persistent asymmetry of this pattern suggests cerebral dysfunction. Excessive anterior 225dysrhythmia can also be seen in some types of cerebral pathology, including meningitis and hypoxic ischemic encephalopathy^17,21,22 (Figure 6.26).

ENCOCHES FRONTALES/FRONTAL SHARP WAVE TRANSIENTS

This pattern consists of 50 to 200 µV, frontal, sharply contoured delta (0.5–0.75 second) waves, often blending with anterior dysrhythmia. There is a small surface-negative phase followed by a large surface-positive phase (resembles a checkmark).²³ Encoches frontales occur between 34 and 44 weeks PMA and are most prominent around 35 weeks PMA. They are rarely seen in neonates by 46 weeks PMA. This pattern is seen most frequently during the transition from active sleep to quiet sleep. Encoches frontales are synchronous, symmetric, and may occur in brief runs (Figure 6.27).

Multifocal Sharp Transients and Focal Sharp Waves^7,8,12

Negative sharp transients in the temporal, central, and centrotemporal regions are common in normal neonates who are near term or older. They typically occur in the setting of a normal background. They tend to occur in isolation but may infrequently be seen in brief, nonpersistent runs, defined as three or more sharp transients in a row. There is no standard definition for pathologic sharp transients. Features that suggest that sharp waves are pathologic include: atypical location (frontal, vertex, and occipital), focality (persistence in one region or hemisphere), and occurrence in long or persistent runs.²⁴ Studies aimed at quantifying negative sharp transients suggest that greater than 11/hour in term and 13/hour in preterm infants during periods of continuous background activity is abnormal. Positive temporal sharp waves can also occur in normal preterm and term infants. Three or fewer positive sharp waves in an hour in the preterm and 1.5 or fewer per hour in the term neonate are considered normal. Of note, all types of transients should be quantified during the most continuous background periods.^25,26

While abnormal sharp transients are seen in neonates with seizures, they can also indicate other types of cerebral pathology. This has been described with frontal and multifocal negative sharp transients.²⁷ For example, a higher density of frontal sharp transients may be seen in neonates with hypoglycemia,²⁸ while excess multifocal sharp transients may be seen in diffuse encephalopathy. Negative sharp transients in other areas, for example, the occipital and midline regions, have unclear clinical significance. Abnormal positive sharp transients in the central and vertex regions are primarily associated with intraventricular hemorrhage or periventricular leukomalacia in preterm neonates.^29–32 They can also be seen in the setting of meningitis, hydrocephalus, and hypoxic ischemic encephalopathy. Positive sharp transients may be associated with poorer outcome³³ (Figure 6.28).

Abnormal Patterns^7,8,12

Dysmaturity

A dysmature neonatal EEG is one in which the features seen are features that would be expected in a neonate that is younger than the stated postmenstrual age. The strict definition is an EEG that would be considered normal for a neonate at least 2 weeks younger than the stated postmenstrual age (Figure 6.29).

Excessive Discontinuity

When the IBI duration is longer or the IBI amplitude is lower than expected for gestational age, the background is described as excessively discontinuous, and this may be indicative of cerebral injury, dysmaturity, or encephalopathy from a variety of causes, including medications and noncerebral insults. For example, prolonged IBI can be associated with high blood ammonia levels due to inborn errors of metabolism.^34,35 The report of an excessively discontinuous tracing in isolation may be of limited utility to the clinical care team in light of the nonspecific nature of this finding. Therefore, information regarding the degree of discontinuity and other salient features like sleep wake cycling, reactivity, and variability is also important (Figures 6.30A–B).

Burst Suppression

The most severe form of background discontinuity is the burst-suppression pattern. This pattern is defined as bursts of paroxysmal high voltage mixed frequency (delta, theta) activity interrupted by periods of background suppression (voltage < 5 µV). The background is invariant and does not respond to internal or external stimuli. The bursts can be low in amplitude (<100 µV), and there can be activity during the periods of suppression as long as it is less than 15 µV in a single electrode or less than 2 seconds in duration. No normal age-specific patterns or graphoelements are seen. Burst suppression is highly suggestive of a poor outcome in both preterm and term neonates and can be seen in the setting of severe encephalopathy from many causes, including hypoxia, stroke, inborn errors of metabolism (i.e., glycine encephalopathy, pyridoxine deficiency), and brain malformations^30,36 (Figure 6.30C).

226Abnormal Voltage Patterns

A low voltage suppressed background is defined as a background that is persistently suppressed <10 µV pp without normal background features. The low voltage suppressed background is invariant and unreactive. Higher-voltage activity (>10 µV pp) lasting less than 2 seconds can occur. As with older children, it is important to consider whether the abnormally low voltage is due to cerebral dysfunction or due to extracerebral factors including high electrode impedance, scalp edema, and/or cephalohematoma (Figure 6.31).

Electrocerebral inactivity (ECI) replaces the term electrocerebral silence and describes the absence of background activity >2 µV pp when reviewed at a sensitivity of 2 µV/mm. There are strict technical requirements that must be followed when using EEG to diagnose electrocerebral inactivity.³⁷ An ECI recording can be used as an ancillary test to support a diagnosis of brain death but is neither necessary nor sufficient for a diagnosis of brain death in neonates.³⁸

Effects of Medications on the Neonatal EEG Background³⁹

Intermittent dosing of phenobarbital and benzodiazepines can cause transient amplitude suppression and background discontinuity; this typically resolves by 12 to 24 hours after administration. Opioids given by continuous infusion or bolus dosing can cause increased IBI duration and can disrupt sleep-wake cycling in ill preterm neonates; the EEG typically returns to baseline 4 to 24 hours after cessation of medication administration. Term neonates receiving opioids can also have disrupted sleep-wake cycling.

Effects of Therapeutic Hypothermia on the Neonatal EEG Background

Therapeutic hypothermia at 33.5 °C to 34.5 °C in neonates with hypoxic ischemic encephalopathy does not significantly change background EEG activity. Similarly, infants undergoing surgical repair of congenital heart disease have no change in EEG activity when cooled to 32 °C to 34 °C; however, deep hypothermia below 30 °C with cardiopulmonary bypass results in changes in the EEG background, including discontinuity.^40,41 EEG background patterns have a similar ability to predict outcome in neonates with hypoxic ischemic encephalopathy, regardless of whether they undergo therapeutic hypothermia; however, in hypothermic neonates, prognosis is more accurate at 72 hours,⁴² whereas the EEG background within 24 hours can predict outcome in normothermic neonates. This suggests that hypothermia to 33.5 °C to 34.5 °C does not alter EEG activity but shifts the time points when EEG background provides accurate prognostication.

Neonatal EEG Reporting⁷

The ACNS guideline has suggested that specific clinical information and background features be included in the neonatal EEG report. Important clinical information includes postmenstrual age, neuroactive medications, and use of hypothermia. Epochs to include are the first hour of the EEG background and epochs when there is a significant change. Important background features to report include the presence or absence of state change. Seizure descriptions should include onset, burden, and resolution. Several approaches to reporting seizure burden are suggested. In patients with status epilepticus, the presence, onset, and resolution should be included. There is little standardization across centers and among centers for reporting aside from what is documented in the ACNS guideline.

Conclusion

Neonatal cEEG is vital to the assessment of critically ill neonates. The visual assessment of the neonatal EEG can feel daunting in light of the many elements of the neonatal EEG, each of which is influenced by gestational age. A systematic approach with a focus on the correlation between the behavioral and EEG states is essential to accurate interpretation.

Only gold members can continue reading. Log In or Register to continue