The Concept of Postconceptional Age

The EEG of newborns is strikingly different from that of older children and adults. In fact, the best known elements of the mature EEG (posterior rhythm, sleep spindles, vertex waves) do not make their first appearance until 6 to 8 weeks after term. In the context of electroencephalography, a newborn’s degree of prematurity is stated in terms of postconceptional age (CA). The CA at birth is equivalent to the gestational age and is usually estimated using the date of the mother’s last menstrual period, but other information such as early fetal ultrasounds and the baby’s physical examination can be used to modify the estimate. By definition, a full-term newborn has a CA of 40 weeks and newborns delivered before 37 weeks are considered premature. Note that a 3-week-old newborn who was delivered at 38 weeks gestational age is considered to have a CA of 41 weeks for the purposes of EEG interpretation. The current CA is derived by adding the gestational age at birth to the current age in weeks (time since birth or “legal age”). One of the underlying assumptions of neonatal electroencephalography is that the expected appearance of a healthy newborn’s EEG is based on its CA. Whether it was born prematurely or not, the EEG is generally assumed to evolve at the same rate whether the baby is inside or outside the womb. Certain pathological processes may, however, interrupt this orderly maturation. Therefore, a normal baby born at 41 weeks CA is generally expected to have an EEG structure similar to that of a normal 5-week-old baby who was at born 36 weeks CA.

From extreme prematurity to term to the postterm period, the appearance of the neonatal EEG evolves dramatically. In fact, on the basis of the various EEG features described here, an experienced neonatal electroencephalographer should be able to estimate the CA of a newborn to within approximately 2 weeks from the appearance of the EEG record. It has been claimed that when the CA estimate suggested by an otherwise normal neonatal EEG differs from the estimate based on the baby’s physical examination, the EEG-based assessment is more likely to be correct. Figures 13-1 and 13-2 show the striking changes in the appearance of the cortical surface between 31 weeks CA and 40 weeks CA (term). It should come as no surprise that the appearance of the EEG evolves rapidly in premature babies.

Figure 13-1 A T2-weighted magnetic resonance imaging scan of the brain of a normal baby at 31 weeks CA shows the relatively smooth appearance of the cortical surface and the rudimentary gyral pattern seen at this gestational age. The small amount of cerebrospinal fluid over the surface of the hemispheres appears white in this sequence. Prematures between 24 and 30 weeks CA have an even less developed cortical folding pattern.

Figure 13-2 A T1-weighted magnetic resonance imaging scan of a term baby (40 weeks CA) shows a significant increase in the complexity of the gyral pattern compared with the 31 week CA brain shown in the previous figure. At term, the complexity of the cortical gyral pattern approaches that of the adult pattern. The cerebrospinal fluid over the brain surface appears dark in this scan sequence. This scan also shows a subgaleal hematoma over the right parietal area (arrow). The white signal encircling the head represents scalp fat rather than bone; this scalp hemorrhage lies outside the bony cranium. Such extra-axial collections may decrease the recorded voltages over affected areas.

Recording Technique

Opinion varies as to whether a full or reduced electrode set should be used for neonatal recordings. Some authors assert that the head is smaller, and therefore it is reasonable to apply fewer electrodes to the smaller head of the newborn. The opposing view holds that if the neonatal brain is conceptualized as a shrunken version of the adult brain, each lobe, gyrus, and cortical circuit is proportionally smaller, and the electric fields of discharges will be correspondingly smaller, requiring the usual (nonreduced) number of electrodes to achieve the same anatomic resolution of electric fields. Our laboratory uses a full complement of electrodes from the 10-20 system in newborns and even in most premature infants; reduced electrode sets are only used for premature infants with the smallest head sizes. Although reduced (double-distance) electrode applications have been shown to record the majority of normal and abnormal EEG activity and may also be better tolerated by the premature infant whose scalp skin is more sensitive, occasionally a highly focal seizure discharge or other highly focal finding may be missed. In addition, difficulties with artifact identification represent a hidden pitfall of the use of sparser electrode arrays. When a deflection is seen in a single channel, denser electrode arrays help determine whether an electric field surrounds the event, increasing or decreasing the chances that it is of cerebral origin as opposed to an electrical artifact.

Additional leads are applied to help assess sleep state; to some extent, a neonatal EEG recording resembles a polysomnogram. The added leads may include a nasal thermistor to measure respirations, ocular leads (one placed just above the outer canthus of one eye and the other just below the outer canthus of the other eye), and a submental electrode to monitor chin muscle (EMG) activity. Additional leads may include a strain gauge placed on the abdomen to record respiratory muscle effort and limb leads to document movements. Notations made by the recording technologist on the EEG record should also carefully document the appearance of the baby. Notations such as “appears asleep,” “has hiccups,” “feeding,” “eyes closed,” or “moving” help the reader assess sleep state and evaluate artifacts (see Figure 13-3).

Figure 13-3 Rhythmic waves seen in the fourth to seventh seconds of this neonatal recording represent patting artifact. Because caretakers often attempt to soothe a crying baby by patting, patting artifact is a common finding in the newborn EEG. This type of rhythmic artifact may, in some cases, mimic an electrographic seizure.

Traditionally, newborn EEGs have been recorded at “half ” paper speed (15 mm/sec). Although this practice may have originally been motivated in part by the urge to save paper on long recordings, the compression of the EEG resulting from slow paper speeds can make it easier to identify some discontinuous or bursting patterns, both normal and pathological. Certain delta patterns are easier to appreciate when displayed at slow paper speeds. For these reasons, slow paper speeds are still preferred by many readers for review of newborn EEG recordings. Ideally, a neonatal EEG record should include all stages of sleep—wakefulness, quiet sleep, and active sleep—which often requires recording times over 1 hour to allow assessment of sleep architecture.

The Low-Voltage Irregular Pattern

As the name implies, this pattern consists of continuous low-voltage irregular (LVI), mixed frequencies, with delta and theta activity most prominent. Voltages generally range from 15 to 35 μV. An example is shown in Figure 13-5. As described later, the LVI pattern is seen during both wakefulness and active sleep. The LVI pattern is not expected to be seen during quiet sleep.

Figure 13-5 An example of a low-voltage intermixed (LVI) pattern is shown, with relatively nondescript mixed frequencies. The LVI EEG pattern is characteristic of both wakefulness and of active sleep. The triangular deflections seen in the top (ocular) channel represent rapid eye movement sleep (dots) indicating that this is an example of active sleep. LUE, left under eye; RAE, right above eye.

The Mixed (M) Pattern

The M pattern is similar to the LVI pattern, but with somewhat higher voltages and a more prominent contribution of slow activity. Continuous mixed frequencies are seen with a mixture of voltages (see Figure 13-6). The M pattern can be seen during any sleep stage. During active sleep, the LVI pattern is most characteristic, but the somewhat higher voltages of the M pattern may also be seen. Similarly, during wakefulness either the LVI or M pattern may be seen. In quiet sleep, the tracé alternant and high-voltage slow (HVS) patterns (described next) are most characteristic, but the M pattern may also be seen. Because it is possible to see the M pattern in any stage of wakefulness or sleep, polysomnographic findings and observed behaviors are key to correct determination of sleep stage.

Figure 13-6 The mixed or “M pattern” is similar to the low-voltage intermixed (LVI) pattern, but with higher voltages. The M pattern may be seen in any of the sleep stages, including wakefulness, quiet sleep, and active sleep. Assignment of sleep state when the M pattern is present depends on other recording parameters such as technologist observations and information from the polysomnographic channels.

The High-Voltage Slow (HVS) Pattern

The HVS pattern is characteristic of quiet sleep; it only rarely makes an appearance in other sleep stages. Like the LVI and M patterns, the HVS pattern consists of continuous, irregular mixed frequencies, but with higher voltages (50–150 μV). Delta frequencies are more prominent (see Figure 13-7). As described below, discontinuous patterns (tracé discontinu and tracé alternant) are the primary patterns of quiet sleep from the earliest post-conceptional stage to 38 weeks CA. As the baby gets closer to term, the tracé alternant pattern is replaced by the HVS pattern.

Figure 13-7 This segment of high-voltage slow (HVS) pattern was obtained during quiet sleep in a term newborn. Voltages are higher than were seen in the previous two patterns, but frequencies remain mixed and the waves are irregular. The HVS pattern is particularly associated with quiet sleep.

The LVI, M, and HVS patterns all consist of continuous irregular, mixed frequencies. The main distinguishing feature among these three continuous patterns is voltage.

The Tracé Discontinu Pattern

The tracé discontinu pattern (French for “discontinuous tracing”) is a pattern of early prematurity, seen primarily at 30 weeks CA and before. As the name implies, tracé discontinu is a highly discontinuous pattern consisting of very high voltage polymorphic bursts, often containing large numbers of sharp features that may even resemble high voltage polyspikes (see Figure 13-8). The dramatic bursts of tracé discontinu are separated by equally dramatic flat periods that may exceed 10 to 20 seconds in length in the most premature babies (see Figure 13-9). Because of its resemblance to burst-suppression, a well-known pathologic pattern in adult EEG, it took some time for neonatal electroencephalographers to confirm that this was a normal pattern of early prematurity.

Figure 13-8 When first encountered, the tracé discontinu pattern may appear highly abnormal to the reader accustomed to interpreting adult EEGs. High-voltage bursts containing large amounts of polymorphic activity, often very sharp as in this example, are seen synchronously in both hemispheres. The bursts are separated by quiet periods of varying duration. LUE, left under eye; RAE, right above eye.

Figure 13-9 In this example of tracé discontinu the interburst periods are particularly lengthy; note the time scale at the bottom of the figure. The plentiful spikes seen within the bisynchronous bursts are considered a normal feature of the tracé discontinu pattern. At the earliest CAs, the flat periods between bursts can be quite lengthy, sometimes exceeding 20 seconds. LUE, left under eye; RAE, right above eye.

The Tracé Alternant Pattern

Tracé alternant (French for “alternating tracing”) is the hallmark pattern of quiet sleep in newborns. Tracé alternant is a discontinuous pattern consisting of bursts of mixed activity lasting 2 to 8 seconds with interspersed flatter periods referred to as “interbursts” lasting 4 to 8 seconds (see Figure 13-10). Generally, the bursts and interbursts are of similar duration. The bursts normally contain a variety of activity, including sharp transient activity and also delta brush activity in more premature babies (described later).

Figure 13-10 The distinctive, discontinuous pattern of tracé alternant is shown. In this example (A), most of the bursts are bisynchronous with bursting activity and suppressions occurring in each hemisphere more or less at the same time. Some amount of asynchrony is noted, however. The lower panel (B) shows the same page of EEG as was shown in Panel A, now with shading marking the approximate beginning and end of each burst. Note each burst contains a fair amount of sharp activity. The regular respirations and lack of eye movements confirm that this is an example of normal quiet sleep.

When tracé alternant makes its first appearance after the 30 weeks CA, the quiet interburst periods are longer and flatter than at later CAs. Also, early on, the bursts show the least amount of synchrony between the two hemispheres. As the baby approaches term, the tracé alternant pattern evolves in three ways. First, the bursts are not as widely separated (the interburst intervals are shorter). Second, the periods between the bursts evolve from being relatively flat showing only small amounts of activity to showing increasing amounts of activity, so much so that as term approaches, it may become difficult to tell where a burst ends and a quiet period begins. Finally, the degree of interhemispheric synchrony of the tracé alternant bursts increases toward term, although it may never reach complete synchrony. The pattern shown in Figure 13-11 has, indeed, achieved complete synchrony, although this does not always occur. Even after term the degree of interhemispheric synchrony of tracé alternant is never required to exceed 75%, meaning that in normal babies, a small amount of asynchrony may always be seen.

Figure 13-11 In this example of tracé alternant the bursts and suppressions are completely synchronous. The time scale has been compressed to make the discontinuous pattern easier to appreciate. Although clearly of lower voltage than the bursts, the periods of suppression contain a fair amount of activity. This and the high degree of synchrony suggest that the baby is near term.

The differences between tracé alternant and tracé discontinu are both qualitative and quantitative. Quantitative differences include longer interburst intervals, more sharp activity within bursts, and near complete synchrony in tracé discontinu compared with tracé alternant. Qualitatively, in tracé discontinu the interburst intervals are expected to be essentially flat, whereas varying amounts of continuous activity are expected during the interburst intervals of tracé alternant. Between 30 and 34 weeks CA, the evolution of tracé discontinu to tracé alternant during quiet sleep occurs on a continuum.

The Evolution of Interhemispheric Synchrony

In contrast to tracé alternant, the tracé discontinu pattern is almost completely synchronous between the hemispheres. This leads to a distinctive sequence in the evolution of interhemispheric synchrony through the weeks of prematurity. In the most premature babies, there is nearly complete interhemispheric synchrony, and the tracé discontinu pattern persists (up to about 30 weeks CA). When the tracé alternant pattern first appears (at approximately 30 weeks CA), the pattern is initially significantly asynchronous. This is followed by a gradual return of interhemispheric synchrony as the tracé alternant matures as the baby approaches term. Therefore the EEG is synchronous in early prematurity, becomes moderately asynchronous in “middle” prematurity, and becomes synchronous again near term.

Active Sleep

During active sleep, the baby is seen to squirm, grimace, and have an agitated appearance, yet the eyes are closed. In fact, the movements may lead an observer to think that the baby is on the verge of waking up. Respirations are irregular, and occasional respiratory pauses may be seen. Rapid eye movements of sleep are seen, both on the eye channels of the EEG and by casual observation of the baby’s eyelids; movements of the corneal bulge can be seen through the baby’s eyelids. The chin EMG lead picks up phasic bursts of muscle activity that correspond to facial muscle movements, such as grimacing or other movements. However, in between facial movements, chin EMG activity is low. The EEG shows an LVI pattern that is similar to what is seen during wakefulness (see Figure 13-12). Although most active sleep stages are typically associated with an LVI pattern, the first period of active sleep occurring as a baby falls asleep may show a somewhat higher voltage EEG pattern compared with later active sleep stages, such as an M pattern.

Figure 13-12 An example of active sleep is shown. Note the low-voltage irregular (LVI) pattern in which overall voltages are low save for examples of superimposed motion artifact. The respiratory (top) channel shows irregular respirations and a brief respiratory pause (arrow), consistent with active sleep. The oculogram (second channel) shows sharp deflections representing horizontal rapid eye movement of sleep (dots), the hallmark of active sleep.

Active sleep in neonates is analogous to REM (dream) sleep in children and adults, although there are two interesting distinctions. First, although older subjects experience a form of paralysis during dream sleep, presumably so that dreams are not physically acted out, as the name implies, babies move actively during active sleep. Second, whereas the first REM sleep stages typically start only after a considerable time asleep in children and adults, newborns enter active sleep as their first sleep stage at the time of transition from wakefulness to sleep. REM sleep at sleep onset is not expected in adults, except in patients with narcolepsy in whom this phenomenon is one of the hallmarks of the syndrome.

Quiet Sleep

The term quiet sleep derives from the quiet appearance of the baby during this sleep stage. Respirations are deeper and regular, and there are few, if any, limb movements. Outwordly, the baby appears to be in a deep sleep state. REMS are not seen (see Figure 13-9). The chin EMG lead, perhaps surprisingly, shows a high level of tonic muscle activity, with comparatively more EMG activity than is seen between body movements in active sleep. After term, quiet sleep evolves into slow-wave sleep.

The EEG pattern seen during quiet sleep before term is the distinctive tracé alternant pattern, a discontinuous pattern over each hemisphere with periods of high-voltage mixed activity followed by periods of relative quiescence. As the baby approaches term, an HVS pattern gradually replaces the tracé alternant pattern during quiet sleep stages. During this transitional period, which occurs during the weeks just before and after term, some babies manifest an HVS pattern at the beginning of a quiet sleep epoch, which may then transition to a tracé alternant pattern with deepening quiet sleep within the same epoch.