INTRODUCTION

The electroencephalogram (EEG) is the single most informative laboratory test in evaluating children with seizures. At the most rudimentary level, it helps in differentiating epileptic from nonepileptic behavior. Because many conditions, including breath-holding spells, movement disorders, syncope, cardiac arrhythmias, sleep disorders, migraine, and various psychiatric syndromes, may mimic epilepsy, EEG findings are often essential in making an accurate diagnosis.

EEG, however, offers much more. Detailed analysis of epileptiform abnormalities may distinguish focal from generalized processes and even more importantly may help characterize epileptic syndromes, which are indispensable cornerstones of rational therapy. EEG allows recognition of subtle and nonconvulsive seizures, can help to identify antiepileptic drug (AED) toxicity, and is useful in selecting patients for AED withdrawal after remission of seizures. Finally, EEG is critical in evaluating patients with medically refractory seizures for focal resective and other surgical procedures.

EEG TECHNIQUE

To record the EEG, a technician places electrodes, usually gold or silver discs, at standard locations on the scalp using collodion adhesive or a conducting paste. Potential differences between pairs of electrodes are then amplified and the net signal from each amplifier is displayed on a monitor (digital EEG machines), or written on paper (historical EEG machines), to provide a graphic record of EEG voltage changes over time. In practice, modern EEG machines display activity from 20 or more channels (1 pair of electrodes = 1 amplifier channel) simultaneously to provide a comprehensive survey of cerebral electrical activity.

Much of the value of EEG lies in determining the spatial distribution of voltage fields on the scalp. To do this, electrodes are grouped in logical arrangements called montages. Montages typically allow comparisons between symmetric areas of the two hemispheres and between parasagittal and temporal areas in the same hemisphere. Most laboratories use a minimum series of standard montages recommended by the American EEG Society (1). In addition to these, special montages may have to be designed to address issues posed by a particular patient. Creative use of rationally designed montages significantly enhances the utility of the EEG.

A technician typically records spontaneous EEG activity for approximately 30 minutes. However, the yield of positive findings is greatly increased by several activating procedures such as hyperventilation, photic stimulation, and sleep. All three are useful in children with suspected seizures or epilepsy.

Hyperventilation

Hyperventilation for at least 3 minutes, preferably 5 minutes if absence seizures are strongly suspected, should be performed whenever possible. Overbreathing can be achieved even with preschool children if the technician incorporates playful strategies such as blowing a pinwheel or “having a race” to see who can breathe faster. The effect of hyperventilation on EEG activity in children is usually dramatic, with high-voltage, generalized, rhythmic slow waves appearing promptly (Figure 12.1). The mechanisms underlying this age-related normal phenomenon are not known, but changes in cerebral blood flow and the neuronal metabolic milieu are probable factors. What is more important, however, is the empiric observation that a brief period of vigorous overbreathing potently activates generalized epileptiform discharges, especially 3-Hz spike–wave activity. Sometimes focal slow-wave activity associated with structural lesions may appear or be accentuated during hyperventilation. The only responses to hyperventilation that can be unambiguously categorized as abnormal are asymmetric changes and epileptiform discharges.

FIGURE 12.1 Effects of hyperventilation in a 5-year-old child. (A) EEG at onset of overbreathing effort; (B) 45 seconds later. There is a moderate buildup of generalized rhythmic slow-wave activity, maximal over the posterior head regions.

Photic Stimulation

Photic stimulation is performed using stroboscopic flashes of high-intensity white light at the rate of 1 to 30 flashes per second. Stimulation is intermittent, with each frequency delivered for 10 to 20 seconds. A normal physiological response is entrainment of EEG activity over the occipital lobes at the flash frequency (photic driving) (Figure 12.2). In some normal children, photic stimulation produces no effect, and in others the photic driving may be asymmetric. A photoparoxysmal response characterized by generalized bursts of irregular spikes, spike–wave discharges, and multiple spike–wave discharges (Figure 12.2B) occurs in some patients with idiopathic generalized epilepsy. However, it also occurs in a significant percentage of normal children without seizures, presumably as an asymptomatic marker of a genetic trait. Doose and Gerken (2) propose a multifactorial mode of inheritance.

Sleep and Sleep Deprivation

Light sleep substantially increases the percentage of EEGs showing epileptiform activity in patients with epilepsy. The occurrences of both focal and generalized discharges are increased in non-rapid eye movement (non-REM) sleep, but rapid eye movement (REM) sleep has a differential effect. Generalized epileptiform activity diminishes markedly or disappears altogether in REM, whereas focal spikes and sharp waves are either unaffected or may actually increase in abundance (3).

Sleep deprivation also activates epileptiform activity independent of its sleep-inducing effect (4,5). Kellaway and Frost (6) have speculated that sleep deprivation and spontaneous or sedated sleep activate abnormalities via different mechanisms.

Scheduling the EEG at the time a child normally naps facilitates sleep recordings. Older children may be partially sleep-deprived with benefit. If children do not sleep spontaneously, they may be given chloral hydrate (30–60 mg/kg) according to appropriate guidelines for conscious sedation.

Special Electrodes

Some cortical areas (mesial temporal, orbital frontal, and interhemispheric regions) are relatively inaccessible to conventional recording electrodes. If one of these areas is suspected of being the epileptogenic focus, supplemental electrode placements can augment standard ones. Thus, surface placement of anterior temporal, mandibular notch, and supraorbital electrodes are useful in documenting inferior frontal and mesiobasal temporal epileptiform discharges (7).

More invasively, sphenoidal electrodes, which are fine wires inserted transcutaneously through the mandibular notch so that the tips lie near the foramen ovale at the base of the skull, are commonly used in monitoring units evaluating adult patients for resective surgery (8). Nasopharyngeal electrodes, which are flexible wires that are inserted through the nose to lie in the posterior pharynx, have been widely used in the past, but subsequent studies have cast doubt on their superiority over surface locations, especially anterior and inferior temporal leads, in recording mesial temporal lobe discharges (9,10).

Invasive electrodes should not be used in children except under special circumstances. Adequate information can usually be obtained using standard scalp electrodes supplemented, as necessary, with additional special placements.

FIGURE 12.2 (A) Normal physiologic response in entrainment of cerebral electrical activity over the occipital regions at the frequency of light stimulation (driving response). (B) Photoparoxysmal response in another child occurring during stroboscopic light stimulation. Bursts of spikes and irregular spike–wave discharges occur during and immediately after the light stimulus. In this case, there were no clinical manifestations during the photoparoxysmal response and no clinical history of photosensitivity.

SPECIFIC EEG FINDINGS

Epileptiform Activity and Epileptogenicity

The principal EEG finding that indicates a susceptibility to epileptic seizures is epileptiform activity, that is, spikes, sharp waves, or spike–wave discharges (Figure 12.3).

Although epileptiform discharges indicate an increased risk of seizures, they vary in the degree of epileptogenicity, that is, the association with active epilepsy. Epileptogenicity is at least partly age-related. In adults, epileptiform discharges have a high association with seizures and occur only rarely in normal individuals. The situation is more complicated in children, largely because of the occurrence of genetic patterns that may not be associated with clinical seizures. Nonetheless, even in children, epileptiform discharges are uncommon in normal individuals, approximating perhaps 2% in the prevalence studies of Petersen and coworkers (11). From another perspective, Trojaborg (12) found that 83% of children whose EEGs contained focal spikes had seizures.

Another variable that relates to epileptogenicity is location of the discharge. The likelihood of epilepsy is highest when epileptiform discharges are multifocal or when they involve the temporal lobe: 76% and 90%, respectively (13). Risk of epilepsy is considerably lower (approximately 50%) when epileptiform activity involves the central–midtemporal (Rolandic) or occipital regions (13).

FIGURE 12.3 (A) Focal left parietal spikes (discharge is at P3 in third and fourth channels) in a 10-month-old boy with simple partial seizures involving his right arm and face. (B) Generalized 4- to 5-Hz spike-wave activity in an 18-year-old male with idiopathic generalized tonic–clonic seizures, precipitated by stimulation.

All epileptiform activity decreases in abundance with age or the passage of time, and less than 10% persists indefinitely (13). In general, there is no useful correlation in individual patients between the amount of epileptiform activity in the EEG and the number and intensity of clinical seizures.

EEG BACKGROUND ABNORMALITIES

Abnormalities other than epileptiform discharges may commonly occur in the EEGs of patients with epilepsy. Nonepileptiform abnormalities can include focal or generalized slow-wave activity (Figure 12.4) or asymmetries of frequency or voltage. Unlike epileptiform discharges, such findings do not provide support for a diagnosis of epilepsy because of their varied clinical associations and because they do not reflect an epileptogenic disturbance of neuronal function. However, they are crucially important in identifying the cause of epilepsy and correctly securing an epilepsy syndrome diagnosis.

USING EEG TO CATEGORIZE PEDIATRIC EPILEPSIES

Importance of Accurate Clinical Information

EEG is most useful when it is combined with accurate clinical information. One measure of a test’s utility is the positive predictive value (PPV), which is the ratio of the true positives to all positives (true positives plus false positives). The positive predictive value is higher when the false positives are low compared to the number of true positives. To illustrate the dependence of the PPV of EEG upon clinical information, let us consider a population of 1,000 children and ask how many of them have epilepsy as determined by EEG testing. If one randomly screened all of them with an EEG and without any knowledge of their clinical history, then the number of true positive EEGs would be equal to the prevalence of epilepsy in the general population (let us say about 0.5%) times the proportion of EEGs showing spikes in people with epilepsy (approximately 60%). The number of false positives could be estimated to be approximately 2% (from various population-based studies). For 1,000 children the ratio would be 3/23 or only 13%! In contrast, if the clinical history were taken into account and we screened only those with a history truly suggestive of a seizure, then the numbers change considerably. Let us say that we have a roughly 50% likelihood that the patients with a positive history actually have epilepsy; then the PPV of EEG becomes PPV = 300/320 or 94%! This may be the single most important take-home message of clinical EEG: the test offers its greatest aid only when it is combined with an accurate clinical history.

FIGURE 12.4 (A) This EEG sample demonstrates continuous focal arrhythmic slow activity in the right central–parietal region (bottom three channels). (B) In contrast, this EEG sample shows excessive generalized arrhythmic slow activity. The EEG of a normal 4-year-old is shown for comparison (C).

Categorizing Epilepsies Using the Interictal EEG

An unconventional but effective starting point for an organization of pediatric epilepsies is the interictal EEG. As shown in Table 12.1, the various patterns encountered in clinical practice may be reduced to five discrete interictal EEG groups. There are two major domains: the organization of the background and the characteristics of the epileptiform activity, most importantly the morphology of the interictal epileptiform discharges. By considering the age of onset one can narrow down the epilepsy syndromes to two or three possibilities, and the predominant seizure type will easily guide one to the final diagnosis. In a minority of cases, most particularly the familial epilepsies, verification of other similarly affected family members is critical for diagnosis. These groupings provide some powerful information: if there are genetic predispositions, they predict the mode of inheritance; they provide information about the general prognosis; they have strong treatment implications; and they could be used as the basis for referral to tertiary epilepsy centers.

TABLE 12.1

Of course this is contrary to the way we as clinicians normally conduct our evaluations. A precise history and physical should always come first. It is the sine qua non portion of the epilepsy evaluation, the basis of all we do, and the most important daily activity of the child neurologist. A thorough history elicited with minimal interruption accompanied by close observation of the child and family allows us to make the interpersonal connections that are critical for trust and healing.

The interictal EEG, however, is a remarkably powerful tool and even though it sits in second place to our clinical assessment, it nevertheless informs us about the underlying pathophysiology in a manner that our naked senses never could. For many years our predecessors have appreciated that the various epilepsy syndromes have different relative contributions of genetic and structural components. These thoughts were codified in the 2010 publication of the International League Against Epilepsy’s classification committee (14). In this chapter we will see how the interictal EEG informs us about these fundamental characteristics. The premise is simple: genetic and structural factors have markedly different EEG signatures that allow the EEG to effectively categorize the epilepsies. Indeed, the EEG findings can be used as an endophenotype to explore the genetic basis of susceptibility to epilepsy (15). This type of epilepsy syndrome organization is highly practical, and simultaneously it reveals some fundamental principles about the cause of the epilepsies. Another remarkable fact is that this can usually be accomplished with a relatively brief sample of the awake and sleep interictal EEG.

1. The Familial Epilepsies (Epilepsies With Frequently Normal Interictal Backgrounds)

One of the five broad categories of EEG features seen in children with epilepsy is a normal tracing. The precise percentage of normal EEGs seen in patients with epilepsy is difficult to determine, but it may be as low as 8% (16). One reason for a repeatedly normal tracing may be a remote location of an epileptogenic lesion—one that does not readily allow for detection using conventional scalp recording electrodes. Normal tracings are also seen in certain distinctive epilepsy syndromes that share common characteristics; they are familial epilepsies that are inherited in an autosomal dominant fashion and the seizures tend to be focal. These may be considered the best examples of familial epilepsies. Why these epilepsies most often present with normal interictal backgrounds is not entirely clear, but certainly the normal background rhythms speak to the absence of cognitive impairment and disability in the vast majority of individuals with these epilepsies. Here, we encounter the first EEG–epilepsy paradox: Even though these epilepsies are strongly genetically determined and spikes in other conditions have a strong genetic component (vide infra), the most conspicuous feature of the background is the lack of interictal epileptiform activity in these familial epilepsies.

There are familial epilepsies for every epoch of life, starting with the neonatal period and continuing into infancy, childhood, and adolescence. These include benign familial neonatal epilepsy, benign familial infantile epilepsy, autosomal dominant nocturnal frontal lobe epilepsy, autosomal dominant epilepsy with auditory features, and other autosomal dominant temporal lobe epilepsies. It is obvious from the titles of these epilepsies that they vary in age of presentation, the brain region commonly involved, and the clinical manifestations. Generally, the combination of the clinical features and family history, along with the mostly normal interictal EEG data, is sufficient to make a diagnosis, but if confirmation is required, genetic testing can help. By and large, the outcome is favorable, although there are published cases of severe phenotypes (17). This topic is artfully and skillfully covered in Chapters 15 and 16.

2. Genetic Generalized Spike–Wave Epilepsies

Individuals with these forms of epilepsy have normal interictal backgrounds with superimposed generalized spike–wave discharges. Seizures will most often be generalized, though focal features may be present. In the EEG, interictal epileptiform discharges will usually be 3 Hz or greater, though, at times, some slower spike–wave activity (circa 2.5 Hz) may be seen, particularly in a younger child. In all cases these occur in the backdrop of a normally developing child. The classic example is a 3 Hz spike–wave activity seen in children with absence epilepsy (Figure 12.5). There may be associated photoparoxysmal responses and some of these epilepsies will show activation of spike–waves with hyperventilation. Spikes may appear more irregular and demonstrate fragmentation in the sleep record (Figure 12.5B). It is not unusual to see focal features (focal slowing and focal spikes). However, these will show lateral shifts from study to study.

As stated, the interictal EEG background is normal, but there may be some important exceptions: rhythmic activity may intermittently punctuate the record. This may be seen as either intermittent rhythmic theta activity, which is often in the biparietal regions (prominent in many cases of myoclonic–atonic epilepsy described by Doose) (Figure 12.6) or as intermittent rhythmic delta, seen in either the occipital or the frontal regions (OIRDA and FIRDA, respectively).

Generalized spike–waves and other paroxysmal features found in the EEGs of individuals with these epilepsies have been known for some time to be inherited in an autosomal dominant fashion with variable penetrance, so that a high proportion of family members of individuals with primary generalized epilepsy will have generalized spike–waves (18,19). The clinical tendency toward epilepsy also has genetic contributors, but these are more complex than the inheritance of the EEG trait. Although the concordance rate for primary generalized epilepsy in twin studies is high (20), the recurrence risks in first-degree relatives of patients with primary generalized epilepsies are much lower than the truly familial epilepsies with monogenic transmission (21). This argues for a more complex polygenic or oligogenic mode of inheritance.

A wide variety of epilepsies are seen within this group and the prognosis is generally very favorable, although some will require treatment for prolonged periods. Associated epilepsies include: myoclonic epilepsy in infancy (MEI), childhood absence epilepsy (CAE), epilepsy with myoclonic–atonic seizures (EMA) (as described by Doose), epilepsy with myoclonic absence, epilepsy with eyelid myoclonia (Jeavons syndrome), juvenile absence epilepsy (JAE), juvenile myoclonic epilepsy (JME) (Figure 12.7), and epilepsy with generalized tonic–clonic seizures alone. Treatment is generally with broad-spectrum agents with the exception of ethosuximide for childhood absence epilepsy. Patients in this category will also usually not require referral to a tertiary center, unless complications arise or special circumstances present themselves. Details regarding these epilepsies are covered in Chapters 17 to 21.

Maturation EEG Trends Seen in Genetic Generalized Epilepsies

Two prominent maturational trends are seen in the interictal epileptiform discharges of children with genetic generalized epilepsies (GGEs). The first relates to the frequency of the discharges. As pointed out in subsequent chapters, the repetition rate of the generalized spike–wave discharges may often be less than 3 Hz in the infantile forms, is precisely 3 Hz in childhood absence epilepsy, and typically is faster than 3 Hz in the epilepsies presenting in the juvenile period. The second relates to the morphology of the discharges. In general, epileptiform activity becomes more distinct, regular, and stereotyped with advancing age. Note how the epileptiform discharges seen in MAE (Figure 12.8) contrast with the very regular and stereotyped discharges seen in the awake tracings of patients with CAE (Figure 12.5).

FIGURE 12.5 (A) Stereotyped 3-Hz spike–wave activity in a 4-year-old child with childhood absence epilepsy. The bottom channel records a test tone (T) to which the child has been trained to respond by pressing a button (R). During the generalized spike–wave activity, the child’s ability to respond is impaired (absence attack). Normal responsiveness returns immediately upon cessation of the discharge. (B) EEG from same child showing alteration in spike–wave morphology during sleep. The epileptiform activity has become fragmented and contains multiple spikes (polyspikes). The well-formed 3-Hz spike–wave complexes seen during wakefulness no longer occur. Such changes, although varying in degree, are typical of generalized epileptiform abnormalities.

3. Self-Limited Epilepsies With Focal Spikes

The epilepsies in this second group show two common features: They have normal interictal backgrounds and they contain highly stereotyped spikes (Figures 12.9 and 12.10). These may be found in a single focus, in homologous regions, or in multiple foci. In some circumstances, the multifocal spikes may appear almost simultaneously. These have been referred to as “clone-like” and they appear to be driven by a primary spike generator, which is located posterior and, often, is of the smallest amplitude (22).

It is not difficult to find the spikes and they are nearly always enhanced in drowsiness and sleep. Importantly, the epileptiform activity evident in sleep is identical to the activity while awake. On some occasions the epileptiform discharges can be elicited by special maneuvers such as temporary loss of visual fixation or finger-pulp tapping. There is a general pattern for the spikes to be more posterior in the younger child and to “move” more anteriorly with age. Note the posterior location of many of the spikes seen in Figure 12.10 as compared with the central location in Figure 12.9. This was described by Gibbs and referred to as “migrating spikes” (23).

FIGURE 12.6 Children with MAE have prominent runs of rhythmic biparietal showing.

These epilepsies have different clinical presentations depending upon the age of presentation and the prominent location of the active focus. Panayiotopoulos uses the apt term “Benign Childhood Seizure Susceptibility Syndrome” or BCSSS for these conditions. They almost invariably have excellent outcome with regard to the cessation of seizures but may be associated with learning, behavior, or attention issues that cause havoc if not properly recognized and treated. For this reason, one might substitute “self-limited” for “benign.” Some prefer the term “epilepsy” to “seizure syndrome,” which yields a shorter term: Self-Limited Epilepsies with Focal Stereotyped Spikes: SEFSS.

The genetics of the EEG trait for the prototype of this category—Rolandic epilepsy (RE)—have been well worked out: Spikes are inherited in an autosomal dominant fashion with variable penetrance according to age (24). A similar conclusion was reached with late occipital onset epilepsy (25). The clinical predisposition to epilepsy, however, appears to have a relatively small genetic component, as substantiated by an analysis of several twin registries (26). Based on this and other information, it has been estimated that the genetic contribution to the clinical susceptibility is small and certainly not confined to just the transmission of the EEG trait (26). This is an EEG–epilepsy paradox: Although centro temporal spikes are seen in all individuals with RE, which has a genetic component, the trait alone cannot explain the genetic contribution. Moreover, the genetic contribution to the clinical susceptibility for epilepsy appears to be small (27).

FIGURE 12.7 Generalized multiple spike–wave (polyspike–wave) discharge followed by a single spike–wave complex in a child with juvenile myoclonic epilepsy.

FIGURE 12.8 The interictal epileptiform discharges in MAE are often brief generalized discharges with an irregular repetition rate.

Well-described SEFSS include Panayiotopoulos syndrome, benign epilepsy with centrotemporal spikes (BECTS) or RE, and late onset occipital epilepsy (Gastaut type) (28).

The spectrum of these disorders is undoubtedly larger than these well-described syndromes. We have seen children in clinical practice who have self-limited epilepsies with stereotyped spikes in other locations, including the frontal and parietal lobes, but these have not been as well characterized in the literature and are apparently not widely recognized. In addition, there can be very severe forms of almost any epilepsy; atypical benign partial epilepsy is one such example for this group. Many highly regarded authorities also consider Landau–Kleffner syndrome (LKS) and continuous spike-and-wave during sleep (CSWS) as extreme versions of these same disorders, although in 25 years of clinical practice this author has not seen a single case transform from ordinary RE to LKS.

Imaging is not required for children with RE, and probably is senseless in those with multifocal stereotyped spikes, a normal background, normal neurological examination, and concordant history for a self-limited epilepsy. The challenging cases are those children with unifocal occipital or frontal stereotyped spikes. Here, it may be prudent to image to exclude the possibility of an underlying focal structural lesion, even if focal slowing and attenuation are not present.

FIGURE 12.9 Stereotyped di- and triphasic sharp wave discharges occur independently in the central regions bilaterally in this child with partial seizures (benign Rolandic epilepsy).

FIGURE 12.10 Highly stereotyped interical epileptiform discharges are seen in patients with Panayiotopoulos syndrome.

Most children with self-limited epilepsies will not require referral to a tertiary center, assuming the correct diagnosis is secured. Prophylactic treatment is generally avoided, unless seizures are particularly bothersome and recurrent. However, in those children with prolonged seizures like Panayiotopoulos syndrome (PS), it would be wise to provide detailed instructions for an emergency plan, including administration of a rescue benzodiazepine for prolonged seizures. The details of these syndromes are covered in Chapters 22 to 24.

4. Epilepsies With Encephalopathy (Epilepsies With Slowed Backgrounds and Multifocal Pleomorphic Spikes)

In most epilepsies associated with an encephalopathy, the EEG background reveals etiologically nonspecific abnormalities. Hundreds of different causes can produce the same EEG appearance, and the EEGs usually share these common features: diffuse background slowing with pleomorphic multifocal epileptiform discharges. Beyond the first year of life, many will also show diffuse spikes or polyspikes that are variable in morphology. If repetitive, these diffuse epileptiform discharges will usually have a slow repetition rate.

Broadly speaking, there are two subtypes of EEGs found in children with encephalopathy and epilepsy. The first group has a continuous background and can be called the epileptogenic encephalopathies, meaning that the same factor or factors that produced the encephalopathy also caused the epilepsy. The second shows some element of discontinuity, often with admixed electrodecrements, and can be termed the epileptic encephalopathies, signifying that the epilepsy per se may be contributing to the encephalopathy. Examples of the latter include early infantile epileptic encephalopathy, West syndrome (Figures 12.11 and 12.12), late infantile epileptic encephalopathy, and Lennox–Gastaut syndrome (Figures 12.13 and 12.14).

There may be various contributions of genetic, metabolic, and structural causes in each individual. When genetic, though, some general rules apply. It is rare to find a familial predisposition to either the EEG features or the epilepsy per se in these patients. There are important genetic causes for the encephalopathies, but for the most part they are de novo mutations, and far less frequently diseases caused by recessive inheritance. Standard autosomal dominant inheritance is very rare. Whole exome scanning has led to the discovery of more patients with oligogenic causes, meaning that two or more gene mutations may act together to produce the precise phenotype. For many epileptogenic encephalopathies it is difficult to define a precise epilepsy syndrome, but further work may lead to the discovery of new electroclinical syndromes within this group.

FIGURE 12.11 Hypsarrhythmia in a 9.5-month-old infant with infantile spasms. There is extremely high-voltage slow-wave activity diffusely, multifocal spikes, and lack of normal organization and patterns.

FIGURE 12.12 Electrodecremental event in a 7-month-old child characterized by transient flattening of background activity (middle of figure). Such events frequently accompany the massive spasms of West syndrome.

FIGURE 12.13 Generalized sharp- and slow-wave complexes (slow spike-and-wave discharges) occurring repetitively at about 2 Hz in a child with Lennox–Gastaut syndrome.

FIGURE 12.14 Tonic seizure during sleep in a child with severe mental retardation and uncontrolled tonic, atonic, and atypical absence seizures. The EEG correlate is an electrodecremental event. Sometimes a low-voltage, 16- to 25-Hz rhythmic discharge (tonic seizure pattern) occurs during this type of clinical event.

4a. Epileptogenic Encephalopathies

The interictal EEG of patients with epileptogenic encephalopathies shows diffuse background slowing with superimposed multifocal pleomorphic spikes. Diffuse spikes may also occur, but these will not appear uniform and will vary in morphology from one complex to the next. Background voltages, even in infancy, are below 300 microvolts. If severe, the backgrounds may lack an anterior to posterior voltage and frequency gradient and sleep architecture may be disrupted. In contrast to the true epileptic encephalopathies, the backgrounds tend to be continuous and electrodecrements are not a frequent feature.

There are few well-recognized syndromes in this category, and one could argue that this grouping contains a collection of specific diseases rather than a list of epilepsy syndromes. It is entirely possible, however, that new electro-clinical syndromes will be described as work advances linking new genetic discoveries with the clinical presentation. Some broad themes may emerge, reflective of the underlying pathophysiology, but for the time being we must await a better characterization of this complex group of patients. It is frustrating work for the even the most skilled pediatric electroencephalographers, because the EEGs by themselves seldom suggest a specific diagnosis. (Notable exceptions to this would be ring chromosome 20, malignant migrating focal seizures of infancy, and Angelman syndrome.) Usually it is the details of the clinical presentation that suggest the underlying cause (eg, Dravet disease, epilepsy in females with mental retardation, etc.) These are covered in more detail in Chapters 25 to 28.

4b. Epileptic Encephalopathies

As mentioned earlier, there is a second subgroup of epilepsies with encephalopathy where the EEGs show some degree of discontinuity, manifested as a discontinuous background, with electrodecrements, or both. The most important reason to recognize this group is the possibility that the epilepsy, per se, can contribute to the encephalopathy; therefore, these conditions have urgent treatment implications. These may rightfully be considered epileptic encephalopathies. Ohtahara conceptualized a spectrum with age-related differences in clinical and EEG manifestations, with a common thread of severe EEG findings, epileptic encephalopathy, and refractory seizures including tonic spasms (29). Examples include the early epileptic encephalopathies (including early myoclonic encephalopathy and early infantile epileptic encephalopathy), West syndrome (Figures 12.11 and 12.12), late infantile epileptic encephalopathy, and Lennox–Gastaut syndrome (Figures 12.13 and 12.14). The additional importance of the electrodecrements is that they indicate a susceptibility to epileptic spasms, myoclonic–tonic and generalized tonic seizures—seizure types that do not respond well to conventional antiseizure agents.

5. Focal Structural Epilepsies

A common cause of refractory focal seizures in children is a focal structural lesion involving the cortical grey mantle, and one of the most common lesions seen at epilepsy surgery is a cortical malformation. Low-grade gliomas and vascular lesions are other important etiologies. Focal structural lesions, regardless of their nature, are often associated with focal slowing, attenuation, or both (Figure 12.4A). In addition, interictal epileptiform discharges tend to be pleomorphic with variations in their precise morphology and topography. This contrasts with the spikes seen in those self-limited epilepsies with focal seizures—in the latter, the spikes tend to be uniform or stereotypical. One way to remember this distinction is to imagine spikes arising from any of a number of locations surrounding a cortical lesion, whereas the self-limited epilepsies without focal structural lesions have an innate driver of the spikes that reproduces the spikes with fidelity from one instance to the other. A high-quality MRI with a careful clinical reading is of course most useful, but a considerable amount of information about the location of the lesion can be gleaned by a careful inspection of the interictal and ictal EEG coupled with an analysis of the ictal semiology. This allows one to focus on a particular region of the MRI, and that specificity helps to uncover subtle cortical lesions.

The ictal semiology in young children has some inherent limitations relating to the ability to precisely localize lesions. The clinical manifestations of seizures in infants and younger children differ markedly from those in older children and adults; the net impact of these differences restricts our ability to precisely localize an ictal discharge based solely on clinical features (30,31). These differences relate, at least in part, to factors intrinsic to the immature brain that bestow unique electrophysiological characteristics, including the underlying normal brain development, the topography of brain metabolism, development of myelinated connections, and properties of ion channels and their associated ion gradients.

As the child matures, the intrinsic properties of brain physiology change and thereby alter the expression of seizures. Gradually, seizures take on characteristics seen in adults. These changes occur in an orderly fashion so that an ontogeny of ictal semiology can be described, just as one can characterize and predict normal child development (32,33). A general understanding of these key differences and some detailed knowledge of the electro-clinical correlation of infantile seizures allows the examiner to ask better questions during the medical interview, aids the evaluation of epilepsy surgery, and increases the chances of making a correct epilepsy syndrome diagnosis.

Clinical Utility of an EEG-Based Categorization System

Even if a specific epilepsy syndrome cannot be secured, the general categories of pediatric epilepsy can provide some useful clinical information (Table 12.1). The first three groupings of EEG patterns/epilepsies occur in the context of a normally developing child with a normal EEG background for age, and either no spikes or stereotyped spike morphology. These children generally do well and usually do not require extensive evaluations and can likely be managed without special resources. There are always exceptions to every rule. All infants and those children who fail to respond to the first one or two selected medications within these groups may benefit from special scrutiny at a tertiary center. Conversely, the fourth and fifth EEG patterns/epilepsies occur in the backdrop of developmental delays or focal neurological deficits, slowed EEG backgrounds, and pleomorphic spikes. These epilepsies raise more concerns and would benefit from prompt evaluations at centers with special expertise in the management of children with epilepsy, including surgical evaluations. Patients in groups 4 and 5 require a careful review of imaging to look for focal structural lesions that might be helped by surgical treatment, as well as for clues to other causes of epilepsy. In addition, patients in group 4 can benefit from metabolic and genetic tests, particularly in those with early onset of epilepsy. The emphasis of this evaluation should be on identifying those conditions that respond to specific treatments such as Glut1DS, pyridoxine-responsive seizures, and creatine deficiency, to name a few. Equally important are the specific treatments to avoid, such as valproate in patients with Alpers, or phenytoin in progressive myoclonus epilepsies.

CONCLUSION

Stay updated, free articles. Join our Telegram channel

Join