04 Normal EEG from infancy to adolescence

There is no such thing as a normal pediatric EEG in the abstract. A reactive 4 Hz posterior rhythm is normal at 4 months and slow at 18 months. Bursts of high-voltage rhythmic theta in drowsiness are normal at 5 years and would be abnormal in an adult. Asynchronous sleep spindles are expected in the first year and would be a finding later. The reader's first act, as Module 2 made a reporting rule, is to fix the age, and the second is to recall what that age should look like.

The table below collects the maturation milestones this module teaches. It is the reference for the sections that follow, and it is the thing to have in mind before opening any child's record.

A caution about the numbers

Sources give slightly different ages for the same feature, because they studied different cohorts with different methods. The 12-month posterior rhythm is 5 to 6 Hz in the Current Practice chapter and 5 to 7 Hz in St. Louis 2016. Treat the table as a guide to what is typical, not as fixed cutoffs, and when a child sits at the edge of a range, say so in the report rather than forcing a verdict. A rhythm one hertz below the typical value in a drowsy or unwell child is a prompt to check the state and look again, not an abnormality on its own.

The posterior dominant rhythm is the single most useful number in a pediatric record, because it summarizes the maturation of the whole background in one measurement, and because its trajectory is well described.

A reactive occipital rhythm of 3.5 to 4.5 Hz is present in about 75% of infants at 3 to 4 months (Current Practice chapter). By 12 months it has reached 5 to 6 Hz in about 70% of infants, or 5 to 7 Hz in the St. Louis 2016 account. By 3 years the mean is 8 Hz in 82% of children, with a range of 7.5 to 9.5 Hz. The mean reaches 9 Hz by age 9 and 10 Hz by age 15 (Current Practice chapter). So the rhythm gains roughly one hertz a year over the first three years and then slows its climb to the adult value through adolescence.

What to measure, and how

Frequency is measured in the fully awake, relaxed child with eyes closed, counting cycles against the time marks rather than estimating by eye, a habit Module 1 insisted on because the time base can mislead. Reactivity is the second measurement: the rhythm should attenuate with eye opening and return with eye closure. A rhythm of the right frequency that does not react is not a normal posterior dominant rhythm. Symmetry is the third: the two occipital regions should show similar frequency and amplitude, and a persistent difference in frequency between sides is a finding even when both are within range for age.

Why it matters at the bedside

A posterior rhythm slow for age is the commonest objective sign of a diffuse disturbance of cortical function, and Module 8 takes up what diffuse slowing suggests. But the finding is only as good as the state in which it was measured. The most frequent reason for a "slow" posterior rhythm in a child is that the child was drowsy, so before writing the number into the impression, find the most alert stretch of the record and measure there. If the child was never fully awake, the report should say the rhythm could not be assessed, not that it was slow.

Sleep is where the pediatric record is richest, and also where it is most often misread. The features appear on a schedule.

Sleep spindles appear at about 2 to 3 months post-term. In the first years they are often asynchronous between the hemispheres, and that asynchrony is normal until about 18 to 24 months (Current Practice chapter; St. Louis 2016). A resident who has learned adult EEG first may read asynchronous spindles in a 1-year-old as a lateralized finding; in an infant they are expected. After about 2 years, persistent asynchrony or absence of spindles on one side becomes a finding.

Vertex waves are seen often by 3 months and reliably by 6 months (Current Practice; St. Louis 2016). In children they can be large and sharp, and they are one of the two drowsy or sleep patterns most often over-read as epileptiform.

Hypnagogic hypersynchrony is the other. It consists of 3 to 5 Hz, 75 to 350 µV bursts of rhythmic activity in drowsiness, occurring from 3 months to 13 years and typically between 4 and 9 years (Kane 2017). The bursts are diffuse, high-voltage and abrupt, which is exactly what a generalized discharge looks like at a glance, and they contain no spikes.

Drowsiness is the danger zone

The over-reading of drowsy patterns is a documented route to the misdiagnosis of epilepsy, as Module 5 shows with the evidence. The defence is procedural: check the state first. Before calling any sharp or rhythmic pattern in a child abnormal, ask whether the child was drowsy or asleep, whether the pattern is one that belongs to that state, and whether it has the features the state predicts, such as vertex waves maximal at the vertex and hypnagogic hypersynchrony without embedded spikes. A pattern that appears only in drowsiness, resolves with arousal, and matches the age-appropriate description is a normal pattern, however dramatic it looks.

Finally, the report should say whether sleep was reached and whether the sleep features were present and symmetric for age. A record without sleep is a weaker negative, and asymmetric or absent sleep features in a child past the age at which they should be synchronous are findings that belong in the impression.

Three awake features appear during childhood, and each is a normal finding that a reader unfamiliar with children can mistake for something else.

Posterior slow waves of youth are slow waves that ride on the alpha rhythm and block with eye opening, like the rhythm they sit on. They are seen typically from 4 to 25 years, and in about 25% of awake children aged 1 to 15 (Kane 2017; Current Practice chapter). Because they are slow, posterior and often asymmetric from moment to moment, they are read as occipital slowing by the unwary. The proof is their behaviour: they are fused with the posterior rhythm, they attenuate when the eyes open, and they return with eye closure. Occipital slowing over a lesion does neither.

Lambda waves are the occipital potentials of visual scanning, and they appear from about 3 to 5 years (St. Louis 2016). They occur with the eyes open while the child looks at something, and they stop when the child looks at a blank surface or closes the eyes. Their sharp contour and occipital location can suggest occipital spikes, but their dependence on visual scanning gives them away.

Photic driving is the occipital response to flashing light, and it is present from about 6 months (St. Louis 2016). It is a normal response, time-locked to the flash frequency or its harmonics, and its absence in an older child is not by itself abnormal. It is important to distinguish it from a photoparoxysmal response, which is an epileptiform discharge provoked by the flash and which Module 9 discusses in the context of the generalized epilepsies.

The common thread

Each of these features is tied to a behaviour: posterior slow waves of youth to eye closure, lambda waves to visual scanning, photic driving to the flash. Each is normal because it appears with the behaviour and disappears without it. When a posterior waveform puzzles you, ask what the child was doing at the time, and check on the video. The state-and-behaviour rule that identified artifact in Module 3 identifies normal variants in the same way.

Activation procedures exist to provoke discharges that a resting record would miss, and in children both procedures have age-specific normal responses that the report must describe correctly.

Hyperventilation is usable from about age 3 (St. Louis 2016); younger children cannot perform it on request. In children who can, marked slowing during hyperventilation is normal, and often dramatic: high-voltage, diffuse, rhythmic slow activity that builds over the procedure and resolves after it. This is a physiologic response and not an abnormality, and calling it one is among the recognized errors that lead to misdiagnosis, as Module 5 documents. What hyperventilation is for is the provocation of generalized spike-and-wave, and Module 2 set out the rule: in an untreated child, 3 minutes of well-performed hyperventilation without 2.5 to 4 Hz spike-wave argues against childhood absence epilepsy. The report therefore says three things: whether hyperventilation was done, whether the effort was adequate, and whether spike-wave appeared. Slowing is described as the expected response.

Photic stimulation produces photic driving from about 6 months (St. Louis 2016): occipital activity time-locked to the flash. Driving is normal and its absence is not by itself a finding. The response that matters is a photoparoxysmal response, an epileptiform discharge provoked by the flash, which is not a normal response at any age and which belongs in the impression. The two are separated by morphology and by behaviour: driving is a rhythmic response that follows the flash frequency and stops when the flash stops; a photoparoxysmal response is spike-and-wave or polyspike-and-wave and may outlast the stimulus.

What the report says

For each procedure: done or not, adequate or not, and the response in standard terms. For a child under 3, state that hyperventilation could not be performed because of age, so the reader knows the negative for absence epilepsy has not been tested. The sign-off for this module, ten normal records from infancy to adolescence with the expected posterior rhythm stated before looking, includes reading the activation sections correctly: a record with marked hyperventilation slowing in a 6-year-old is still a normal record.