Blume's Atlas of Pediatric and Adult Electroencephalography, 1st Edition

Chapter 7

Role of the EEG in Some Pediatric Neurological Problems

Febrile Seizures

A febrile motor seizure occurs with a temperature above 38.4°C that has been present for less than 24 h. The semiology is usually generalized and symmetrical, although one side may predominate. A focal seizure does not fall under this definition. Febrile seizures (FSs) are classified as simple if they are generalized, last less than 15 min, and occur only once in 24 h. Febrile seizures with features outside of this definition are classified as complex.

In their sequential study of patients with simple febrile seizures, Frantzen et al. (1968) found prominent delta activity in one third of the patients during the first week after the convulsion. This excessive delta activity (excessive, as some delta activity is normal in this age group) was either diffuse and maximal posteriorly or confined to the posterior head regions. In most instances it predominated on one side. Delta appeared most commonly and was most marked when (a) the duration of the convulsion exceeded 30 min or had evidence of unilateral predominance, (b) the temperature exceeded 38.4°C, (c) the child was ill for more than 36 hours prior to the convulsion, and (d) diarrhea and vomiting had occurred. In almost all patients, the excess delta activity disappears within 1 week of a simple FS. Given the heterogeneous factors that accentuate such delta activity, it is not surprising that Frantzen and colleagues found that such acute electroencephalographic (EEG) slowing carried no prognostic implications.

The concept that an EEG taken beyond the acute phase is always normal is wrong. It is true that, in most patients, the EEG reverts to normal within 1 week once the delta has receded. However, Alvarez et al. (1980) found bisynchronous spike–waves in early sleep more commonly among patients with FSs than in controls. Forty-two (19%) of the 218 patients of Frantzen et al. (1968) had generalized spike–waves, either at rest or during photic stimulation. These children were slightly older than most when the FSs began; spike–waves appeared more commonly in recordings performed after age 4 years. The incidence of spike–waves was twice as high among those with a family history of epilepsy (Frantzen et al., 1970). Spike–waves appeared very rarely in the acute phase.

The most significant finding of the Frantzen study is that neither generalized spike–waves nor focal spikes served to predict recurrence of FSs or the later development of nonfebrile seizures. Studies by Laplane and Salbreux (1963) and Aicardi and Chevrie (1971, 1973) have indicated that generalized 3-per-second spike waves occurring at age 3 years or less are associated almost exclusively with FSs and/or generalized myoclonic attacks. Aicardi and Chevrie never found generalized spike–waves at less than age 2 years among patients with febrile seizures. Therefore, they felt that the presence of generalized spike–waves at this age may herald myoclonic epilepsy.

The factor of quantity of spike–wave discharges is one unstudied aspect that might help to assess their significance. Although the sampling effect of a limited recording hinders this approach, prolonged recordings may provide an answer.

Clinical Correlations

Three mechanisms may predispose to seizures with a fever. The classic FS represents a genetically determined susceptibility to generalized motor seizures occurring only with fever. A second group of patients seize with fever because of a cerebral lesion occurring either before or during the febrile episode. A third group consists of children who have a chronic generalized epileptic condition that first becomes evident as a motor seizure during a febrile episode.

Complicated FSs are those that last more than 15 minutes, are unilateral or focal, or are repeated within a single febrile episode. Such attacks tend to occur in patients whose neurological development before the febrile attack was already abnormal and usually are associated with a


higher risk of later epilepsy. Conversely, a single, brief, generalized FS has a relatively favorable prognosis. However, in practice, classifying each episode into either the simple or the complicated category may be clinically difficult, as may be the determination of prognosis. Therefore EEG may help to categorize the mechanism of the FS. EEGs obtained less than 1 week after a febrile convulsion may show various quantities of delta activity appearing either diffusely or posteriorly, with the quantity depending on the duration of the febrile convulsion and the interval between its termination and the EEG recording. Such bilateral delta activity would fail to reveal the febrile convulsion mechanism. A postictal EEG with regionally accentuated delta or focal spikes would suggest that the seizure with a fever represented the second and third categories outlined previously—that is, a convulsion secondary to a previous or current central nervous system insult. Regional, hemispheric, or even diffuse arrhythmic, very low frequency delta activity with loss of regional background activity should raise suspicion of an abscess or cerebral vein thrombosis (Kooi et al., 1978).

EEG could be valuable in the emergent situation for any patient who fails to regain consciousness within a reasonable time after the apparent end of a FS in order to exclude the possibility of continuing seizure activity.

EEGs are of limited clinical value in a patient with a simple FS. Clinical judgment would be required as to whether an EEG could help unravel the mechanism of a more complicated attack.

Hemiconvulsion–Hemiplegia–Epilepsy Syndrome

The hemiconvulsion–hemiplegia–epilepsy syndrome described by Gastaut et al. (1960) consists of a unilateral or predominantly unilateral prolonged motor seizure, a postictal hemiplegia that may or may not persist, and a chronic focal epileptic seizure disorder, either as dyscognitive partial seizures from the implicated temporal lobe or focal motor and possible secondarily generalized seizures. The young child is often febrile at the onset of this convulsive status epilepticus. Thus, a prolonged, predominantly unilateral febrile convulsion raises the possibility of this syndrome. Postictally, high-voltage, 1 to 2 Hz delta activity may be seen bilaterally, with emphasis in the implicated hemisphere. This EEG abnormality may persist in a less prominent form for several years. Multifocal spikes chronically appear independently in either hemisphere but principally over the clinically implicated hemisphere. Secondarily generalized spike–waves also are a feature. Neuroimaging may disclose a preceding focal abnormality.

First Seizure

The risk of recurrent afebrile seizures after a first unprovoked seizure in children varies from 42% to 52% (Camfield et al., 1985; Shinnar et al., 1990). Camfield's group found high recurrence rates among those with abnormal neurological examinations, focal spikes on EEG, and dyscognitive seizures. Among children with an idiopathic first seizure, the study byShinnar et al. (1990) found the EEG to best predict recurrence. Abnormal EEGs were associated with cumulative risks of 41%, 54%, and 56% at 12, 24, and 36 months, respectively, as compared with 15%, 23%, and 26% among children with normal EEGs. If epileptiform discharges do not appear clearly on initial recordings, subsequent recordings with sleep may disclose or define spikes more clearly (Carpay et al., 1997; Frost et al., 1991).

Last Seizure?

Prognosis for seizure control hinges primarily upon that of the associated epilepsy. In helping to define that epilepsy, the EEG is useful in forecasting outcome. However, the severity of epilepsy varies among patients with epilepsy syndromes. For those syndromes in which the quantity of spikes or spike–waves reflects seizure incidence, the EEG may influence decisions about medication reduction. Spike–waves and absence seizures exemplify such a correlation (Miller & Blume, 1993; Braathen & Melander, 1997).

Myoclonic Epilepsies of Infancy and Early Childhood

For a child with generalized myoclonic seizures, the clinician wishes to determine the nature of any syndrome that they may represent and thus its prognosis. A major question is whether these myoclonic attacks represent a progressive myoclonic epilepsy, a severe nonprogressive generalized epilepsy, or a more benign epileptic condition.

There are electrographic distinctions between these entities. A slow background rhythm that is not due to drowsiness or medication and the lack of normal sleep patterns are nonepileptiform features of the more


intractable epilepsies such as the Lennox–Gastaut syndrome. More benign myoclonic epilepsies may have a normal alpha rhythm, and the only background abnormality would be occasional “projected” rhythmic delta waves diffusely.

Epileptiform potentials also help distinguish these entities. The more benign myoclonic seizures are associated with occasional spike–waves and polyspike–waves repeating at 2.5 to 4 Hz, even though some more slowly repeating spike–waves may be intermixed. Uncontrolled myoclonic epilepsies may temporarily have slow spike–wave (SSW) discharges, and these could revert to the faster (more than 2.5 Hz) spike–waves with therapy. Spike–waves of the more severe myoclonic epilepsies may be markedly increased or precipitated by photic stimulation. In non-REM sleep, such spike–waves may become polyspike–waves or even brief runs of diffusely distributed polyspikes.

In contrast, the Lennox–Gastaut EEG has SSWs whose repetition rate is usually less than 2.5 Hz. As described in many places elsewhere, these bilaterally synchronous or shifting discharges occupy high percentages of awake and light sleep recordings. Such discharges are almost never precipitated by photic stimulation. Runs of diffusely distributed polyspikes may also appear in non-REM sleep. Brief tonic seizures may be clinically difficult to distinguish from myoclonic attacks; polygraphic recordings may help. The electromyographic (EMG) component would clearly exceed 100 ms in the tonic form, whereas it would be less than 100 ms in myoclonic attacks. Similarly, tonic seizures would be accompanied by a diffuse attenuation of background activity (an electrodecremental event) or high-frequency rhythmic waves either focally or diffusely (Blume, 1982; Gastaut & Broughton, 1972).

During myoclonic attacks of either form, bilaterally synchronous single or very brief multiple spikes may occur. These bear a general relationship to the peripheral myoclonic jerks, but the exact timing mechanism is variable.

Atonic seizures may accompany any of the generalized epilepsies of childhood and are associated with spike–waves, polyspike–waves, fast rhythmic waves (epileptic recruiting rhythm), or SSWs. Absence attacks may also occur. EEGs with benign myoclonic seizures may more resemble typical absence seizures with an abrupt clinical and EEG onset and termination and without associated features such as tonic events. In contrast, those associated with the Lennox–Gastaut syndrome have a more gradual onset and offset with tonic phenomena, automatisms, and autonomic features; their relationship to SSWs is less precise.

EEG features suggesting a progressive myoclonic epilepsy are reviewed under “Progressive Myoclonic Epilepsies” later in this chapter.

Evaluation of Candidates for Epilepsy Surgery

Congruence of several lines of data for the region of epileptogenesis forms the principal consideration in assessing children with intractable focal epilepsy for surgery. Such data include semiology of seizures, interictal and ictal EEG, neuroimaging, and assessment of any focal neuropsychological deficits. Much of this assessment occurs over the several months while medical intractability gradually becomes evident.

During this period, several outpatient EEGs should be performed to determine whether spikes arise principally from one region. Two studies illustrate the localizing value for epileptogenesis of interictal spikes in children. We studied the correlative value of the most active focal spikes to ultimate surgery site in 48 patients undergoing cortical resective epilepsy surgery at age 16 years or less (Blume & Kaibara, 1991). The most active focal spikes correlated with the surgery site in 32 (67%) of the 48 patients. This correlation was highest for temporal lobe resections and lowest for resections in the frontal lobe. Fortunately, 10 (91%) of the 11 patients whose most active spikes resided in an adjacent lobe or were simply ipsilateral to the surgery site had a radiologically demonstrable lesion at the surgical site. Again in this group of 48 patients, 30 had both a clearly defined seizure onset and predominant spike focus. These sites were congruent in 21 (70%) of the 30 patients, whereas the seizure onset was at least ipsilateral to the most active spikes in 7 (23%). No patient had a most active spike focus that falsely lateralized seizure origin.

Of the 48 patients, 20 (42%) demonstrated focal or focally accentuated diffuse delta over the surgery site. In 9 (19%) additional patients, delta appeared either over an adjacent lobe or was simply ipsilateral to the ultimate surgery. In no instance did the most prominent regional delta appear contralateral to the operative site.

In a separate series of 14 children who underwent effective temporal lobectomy for intractable seizures, most active spikes arose from that lobe in 13 (93%) and never falsely lateralized epileptogenesis (Blume et al., 1997). However, temporal spikes tend to appear less commonly in children less than 6 years old. Seizures with an identifiable focal origin arose from the later-resected temporal lobe in 7 (50%); none was falsely lateralizing. Focal delta activity also appeared congruently. Both of these series indicate that, in practice, persistent focal scalp EEG spikes reliably identify seizure origin in most patients, even better than do recorded seizures (Dinner et al., 1984).

Nonetheless, children whose seizures spread rapidly or arise from inferior or mesial–cortical surfaces usually require subdural recordings to clarify the ictal mechanism(s).



Neocortical Malformations

Several EEG phenomena may be associated with neocortical malformations of the brain, including a paucity of EEG activity either focally or diffusely, monorhythmic theta, or diffuse or focal delta activity. Patients who have epilepsy may show multifocal spikes that extend beyond the lobe of the malformation (Palmini et al., 1991a), but these are usually principally seen in the region of the malformation. Focal spikes are very abundant (Raymond et al., 1995) and may arise from the lesion itself as well as from the dysplasia-normal brain interface (Palmini et al., 1991b; Pathak & Blume, 1997). Such discharges may be electropositive (Otsubo et al., 1997). Otherwise normal patients whose intractable focal seizures arise from such lesions are increasingly being identified.

Nonepileptic Events and Video EEG

Psychogenic events and some paroxysmal movement disorders may resemble epileptic seizures in children and adults. A complete event description, clinical neurological evaluation, and outpatient EEG accurately identify the nature of these events in most instances. Synchronized video–EEG monitoring, on either an outpatient or inpatient basis, may establish the diagnosis in the remainder. Some patients with psychogenic events also may have: a) epileptic seizures, b) interictal EEGs containing spikes, or c) both may obtain (Donat & Wright, 1990; Metrick et al., 1991). The value of this procedure is limited by the lack of EEG change in many simple partial seizures and those arising from regions remote from scalp electrodes. Muscle movement and electrode artifacts, which are more common during prolonged recordings, may obscure data.


The sequential EEG events that occur in fortuitously recorded syncope are loss of alpha, a brief period of low-voltage beta, rapidly augmenting diffuse theta, then delta activity followed by transient electrocerebral inactivity, and then progressive recovery. Nonetheless, diagnosis of syncope does not depend on demonstration of this EEG sequence, as was thought in the past.

The danger of performing an EEG in patients with syncope is that some irrelevant anomaly might be disclosed. Therefore, the purpose of the recording should be thoroughly communicated to the patient and parents before the fact. When myoclonic or tonic movements are a prominent or prolonged feature of the syncopal attack (brief myoclonies are common), the clinician may be justified in wondering if a generalized epileptic condition is so represented. However, in almost all cases the description establishes the diagnosis and an EEG is unnecessary.

Acutely Occurring Disorders


Normal background rhythms are replaced by rhythmic and arrhythmic excess delta activity. Such delta may be diffuse, diffuse with transient regional accentuation, or diffuse with consistent regional accentuation. These abnormalities correlate well with the neurological state of the patient, but coincident metabolic and/or electrolytic derangements may contribute to their severity. Resolution of the EEG changes usually parallels clinical recovery, but the EEG may remain abnormal for weeks thereafter (Saunders & Westmoreland, 1979).

There is no convincing evidence that the EEG findings during the acute phase add significantly to clinical data in assessing the prognosis for recovery from encephalitis. For example, prominent diffuse delta activity may resolve completely along with clinical resolution.

However, there are two clinical situations in which the EEG may be useful in caring for a patient with encephalitis. In patients whose neurological function is altered markedly by encephalitis or meningoencephalitis, epileptic seizures may have unusual clinical manifestations or may be clinically undetectable. The EEG can detect such events and can monitor the efficacy of anticonvulsant therapy. As the timing of such seizures may be haphazard, adequate recording time is requisite for this type of evaluation.

In the clinical context of viral encephalitis, the EEG may help to determine whether herpes simplex virus is the etiological agent. Periodic sharp waves repeating every 0.5 to 4 s have been described by several authors in association with herpes simplex encephalitis (Smith et al., 1975; Upton & Gumpert, 1970). Such sharp waves may be diffuse or temporal–frontal and may be unilateral or bilateral. Repetitive sharp waves have not been described in all reported cases of herpes simplex encephalitis, but Upton and Gumpert (1970) emphasize the occasional need for frequent, even daily, recordings to reveal their presence. These sharp waves remain for varying periods over the first 2 weeks of the illness and rarely extend beyond the third week. They are always accompanied by diffuse or


temporally accentuated excess delta activity. This phenomenon is certainly more characteristic of herpes simplex encephalitis than other viral encephalitides. Unfortunately, their specificity for herpes is limited by their resemblance to the periodic lateralized epileptiform discharges that may be found in a wide variety of acute neurological disorders (Kooi et al., 1978).


The acute EEG changes associated with meningitis resemble those of diffuse encephalitis: diffuse excess delta and theta. However, they are less severe and may partially represent associated metabolic derangements or medication effects. Such alterations tend to be more severe in children than in adults. In this clinical context, persistent, very slow arrhythmic focal delta activity should raise the clinical suspicion of a complicating abscess or cortical vein thrombosis.


Several mechanisms may produce acute-appearing focal, multifocal, or diffuse EEG abnormalities after trauma. The most obvious of these is direct trauma to the brain. Multifocal EEG abnormalities would ensue. Carotid artery trauma may lead to a dissection-related stroke; EEG abnormalities confined to one hemisphere would suggest this. Chest injury may produce hypoxic or ischemic encephalopathy; diffuse EEG changes (vs. multifocal) would obtain. Finally, long bone fractures rarely produce fat embolism with multifocal EEG abnormalities.

As with adults, the degree of EEG abnormality after trauma in children is in proportion to the severity of the injury. However, the magnitude of the EEG abnormality usually exceeds that in the adult. Even a mild head injury in a child may be associated with prominent EEG changes. Thus, prominent EEG changes do not necessarily connote irreversible brain injury.

The most striking single abnormality in the acute phase is excess delta activity located principally posteriorly (Frantzen et al., 1958; Silverman, 1962). The younger the child, the lower is the frequency of this delta. Such posterior delta appears to be an acute phenomenon. It declines rapidly after the second week postinjury. Regional delta in other areas disappeared more slowly in the study by Silverman. Trauma-related EEG changes tend to persist longer in children than in adults. Although such changes last longer after severe head injuries, there is considerable intersubject variability in this respect.

Because a head injury involves both direct and contrecoup mechanisms, clinically inapparent dysfunction might be revealed by EEG. For example, hemispheric abnormalities ipsilateral to hemiplegia may occur.

In assessing the effects of head injury on the EEG, keep in mind the possibility of a preexisting abnormality. For example, well-formed, 3-per-second, bisynchronous spike–wave discharges are not produced acutely by trauma (Kellaway, 1955). This underlines the importance of obtaining an EEG shortly after the head injury. Periodic follow-up EEGs are needed to assure resolution of any persisting abnormality, particularly if it is clinically silent.

Following one or a series of trauma-related abnormal EEGs, more than a single normal recording is necessary before one can state that the EEG has returned permanently to normal. Such recordings should include sleep, which may reveal abnormalities not seen during wakefulness.


Clinical evaluation of comatose patients involves examination of functions primarily mediated by the brainstem. Therefore EEG, whose data reflect mainly cortical function, provides an often useful adjunct to the clinical assessment.

Several EEG phenomena appear in comatose conditions. The most common is diffuse persistent excess delta and theta activity. Lack of attenuation or other alteration of this activity by afferent stimuli indicates deep coma. The occurrence of triphasic waves in association with a depressed level of consciousness indicates a metabolically induced comatose condition (Bickford & Butt, 1955; Sundaram & Blume, 1987), but triphasic waves rarely occur in children. Periodic lateralizing epileptiform discharges may reflect superimposed regional abnormalities (Chatrian et al., 1964).

When recurrent seizures complicate the situation, the effectiveness of anticonvulsant treatment can be monitored by assessing the abundance of clinical and subclinical electrographic seizures and the quantity of spikes.

Burst-suppression activity may appear in deep coma: bursts or brief runs of intermixed theta, delta, and spikes are separated by equal or longer periods of relative or complete inactivity, either diffusely or regionally. In other situations, diffuse, nonreactive sinusoidal patterns in the theta or alpha range have been described by several authors (see Bauer [1987] for review).

The prognosis of any of these patterns depends on the etiology of the comatose condition, its duration, and the direction in which sequential EEG recordings evolve. Thus if anesthetics or other central nervous


system depressants have been used, the value of EEG patterns for prognosis is minimal. Metabolic and toxic states usually have a better prognosis than structural or anoxic encephalopathies for a given EEG picture. Within this context, prognostically favorable signs are EEG reactivity to exogenous stimuli, spontaneous variability, and normal sleep potentials. The following suggest an unfavorable outcome: lack of reactivity to afferent stimuli, the burst-suppression pattern, monorhythmic alpha or theta frequencies, a very low voltage EEG, or electrocerebral inactivity.

Pampiglione and Harden (1968) performed EEGs within the first 12 h of cardiac arrest in children aged 1 day to 10 years. All 61 children whose EEGs contained at least some features appropriate to age recovered rapidly from the anoxic episode. A few whose tracings developed some normal features a few hours after initially showing continuous arrhythmic delta also did well. All 10 patients whose EEGs showed continuous delta for many hours without any normal features had unfavorable outcomes: 9 died and 1 remained decerebrate. All 27 patients with burst suppression and the 7 patients with electrocerebral inactivity died. After cardiac arrest, electrocerebral activity never recovered if it remained absent for 2 to 3 h.

Seshia et al. (1979) studied initial EEGs taken in children 12 to 24 h after cardiac arrest. Of their 24 patients, 9 had diffuse, high-voltage delta activity. The EEGs of 4 of these 9 patients progressed to burst suppression or electrocerebral inactivity; all died; 3 others remained handicapped. The 2 patients whose EEGs progressed to normal were neurologically normal at discharge. All 3 patients with only diffuse low-voltage activity progressed unfavorably: 2 died and 1 remained severely handicapped. Those patients with burst suppression (2 patients), diffuse alpha-like activity while comatose (1 patient), and electrocerebral inactivity (14 patients) died.

Combining these data creates the following summary of EEG findings after cardiac arrest in childhood and their prognoses:

  1. When any normal activity is present, the outlook is favorable.
  2. A tracing dominated by high-voltage delta activity requires follow-up recordings before a prognosis can be established.
  3. Very low voltage EEGs convey an unfavorable outlook.
  4. All patients whose EEGs show electrocerebral inactivity persisting for 2 to 3 h or burst suppression succumb unless the EEG is confounded by medically induced coma.

Determination of Irreversible Coma without EEG

It is hoped that the concept of irreversible coma will replace that of brain death in approaching this type of situation. Mohandas and Chou (1971) established criteria that would determine beyond a reasonable doubt the state of irreversible damage to the brainstem: (a) a known, irreparable intracranial lesion; (b) no spontaneous movements; (c) apnea; (d) no brainstem reflexes; and (e) no change over 12 h. No child or adult satisfying such criteria has ever survived (Jorgenson, 1981; Moshe & Alvarez, 1986; Pallis, 1983; Rowland et al., 1983; Robinson, 1981; Tomlin et al., 1981). Bobele et al. (1993) found clinical examination to predict death more reliably in newborns and infants than either the EEG or radionuclide cerebral perfusion scans (see Young [1998] for further discussion).

With respect to the EEG, even apparent electrocerebral inactivity cannot be equated with total cessation of cortical function. A limiting factor is machine noise, from which cerebrally originating potentials of less than 2 µV cannot be distinguished. Ashwall and Schneider (1979) showed the presence of EEG activity in five patients up to 30 months of age who fulfilled other criteria for brain death; none of these patients survived. Blume et al. (1995) found in adults that very low voltage EEG (20 µV or less) carries the same prognosis as electrocerebral inactivity.

The EEG could be employed as an ancillary test in situations where full evaluation of brainstem function is not possible from a practical standpoint. This would occur when trauma to structures reflecting brainstem function (e.g., cranial nerves) had occurred. Even in this circumstance, EEG data would not necessarily assume primary importance but would be considered along with other data in arriving at a clinical decision of irreversible coma.

If one accepts the concept of irreversible coma instead of brain death, demonstration of complete cortical electrical inactivity may not be required. Reactivity of any EEG pattern to afferent stimuli would assume paramount importance. Technical requirements can be found in many publications, particularly that by Bauer (1987).

Chronic Nonprogressive Disorders


Clinical judgment is probably more reliable than the EEG in distinguishing headaches representing progressive lesions such as tumors or vascular malformations from migraine and tension headaches. The matter has never been studied systematically in children and adolescents.

The difficulty lies in the high percentages of EEG abnormalities found in children with migraine, ranging from 44% (Froelich et al., 1960) to 73%


(Prensky & Sommer, 1979). A wide range of abnormalities is encountered, including spikes and sharp waves, bursts of theta, and diffuse or focal delta activity. Focal delta may occur in association with hemisensory and/or hemiplegic migraine, migraine with a visual aura, or uncomplicated migraine. The EEG slowing may persist for 1 to 2 weeks after the episode.

It is equally doubtful that the EEG exceeds clinical judgment in separating migraine from epilepsy in view of the moderate incidence of epileptiform potentials seen in childhood migraine (22% [Froelich et al., 1960] and 47% [Prensky & Sommer, 1979]). In some instances, the conditions are entwined, as in the syndrome of basilar migraine, visual phenomena, seizures, and occipital spikes (Camfield et al., 1978; Gastaut & Zifkin, 1987).

The foregoing indicates that it is unnecessary to include the EEG in the clinical assessment of the majority of children with headaches. In complex cases, the EEG may assist in a judgment that remains primarily clinical.

Cerebral Palsy

Gibbs and Gibbs (1964) carried out the most comprehensive study of EEG changes in association with cerebral palsy. The following data are taken from their report unless indicated otherwise.

Of all forms of cerebral palsy, patients with hemiplegia have the highest incidence of EEG changes (90%), followed closely by quadriplegics (85%). As would be expected from the more deeply seated pathology, abnormal EEGs are less common with paraplegia (70%) and athetosis (50%). The incidence of seizures among these groups closely parallels that of EEG abnormalities.

Of patients with cerebral palsy, 14% have abnormally low-voltage awake background activity. V waves and spindles are reduced or absent in 5% to 45%, the lower figure for athetosis and the higher for quadriplegia. Asymmetries of awake and sleep potentials occur most commonly among hemiplegic patients, in whom the side of lower voltage always corresponds to the clinically implicated side for sleep potentials and usually corresponds for awake potentials.

Taking all age groups together, the most common single abnormality is multiple independent spike foci. Although the most actively spiking area on scalp recordings may relate to the side implicated by a hemiplegia, in some cases it may be over the healthier hemisphere (Rasmussen, 1975). This is particularly the case when extensive destruction of one hemisphere has occurred. In this situation, the voltage of bilaterally synchronous discharges may also be greatest over the relatively intact side. The background activity over the involved hemisphere will be reduced.

Hypsarrhythmia is the most common type of abnormality in those less than 1 year of age. The occipital lobe is the most common focus of spikes up to 10 years of age, when the temporal lobe becomes the principal site. However, no single region strongly predominates at any age.


Children with autism display several unusual EEG features (Dorenbaum et al., 1987). Rossi et al. (1995) found multifocal spikes including rolandic spikes in EEGs of these patients. An EEG including sleep or an overnight video-EEG could be performed seeking subclinical seizures.

Angelman Syndrome

According to Laan et al. (1997), the most typical EEG findings are high-voltage rhythmic triphasic delta waves, maximal frontally and occurring intermittently or continuously. In contrast, Boyd et al. (1988) described high-voltage bursts of posterior-dominant, 3-per-second rhythmic waves occasionally with intermingled spikes.

Degenerative Disorders

The differential diagnosis of a patient with apparent degenerative disease of the central nervous system will be determined largely by age, symptoms, course, neurological examination, and family history. Within this context, the EEG may help to narrow the range of diagnostic possibilities. Additionally, its findings may spark the suspicion of a degenerative disorder in an otherwise healthy patient with a single symptom, such as a seizure.

Although the nature of the degenerative process heavily influences the EEG findings, other factors play significant roles. These include age, stage of disease, and the presence of any systemic condition complicating the disease. Such factors should be considered whenever the EEG findings are at variance with the clinical diagnosis.

A slowing and then loss of normal background rhythms combined with excess delta and theta occur in all established degenerative conditions involving the cerebral cortex and/or the underlying white matter.

However, Gloor et al. (1968) were able to correlate the pattern of distribution of the lesions with certain other EEG characteristics. The following principles derive from their comparative study and from reports of individual disease entities as described below. Bilaterally synchronous


paroxysms occur only with cortical and subcortical gray matter diseases. These discharges may consist of bilaterally synchronous slow waves, spike–wave complexes, SSW complexes, or focal spikes. The morphology of this paroxysmal activity is probably more age than disease-dependent within the gray matter disorders. The spike–wave discharges may resemble those of idiopathic epilepsy, but they are usually more abundant.

In contrast, diffuse high-voltage arrhythmic delta activity dominates the EEG of white matter diseases. Both features are seen in association with diseases, such as subacute sclerosing panencephalitis, which involve both gray and white matter. Focal and multifocal epileptiform potentials appear in both gray and white matter diseases but are more abundant with the former.

These distinctions apply principally to stages of these diseases at which the EEG findings are maximally expressed. Early EEGs may be normal or minimally and nonspecifically altered. Near-terminal recordings are characterized by low-voltage arrhythmic waves.

Gray Matter Disorders

Progressive Myoclonus Epilepsies

Slowing and subsequent loss of normal background rhythms and abundant generalized or multifocal spikes and spike–waves occurring spontaneously and evoked by visual stimuli characterize EEGs of these epilepsies (Berkovic et al., 1986). More common rather than distinctive EEG features will be found in the following brief descriptions of selected conditions. Thus, the aforementioned EEG abnormalities signal the possibility of a progressive myoclonic epilepsy more than the identification of the individual disease.

GM2 Gangliosidoses

The earliest change consists of bursts and runs of high-voltage theta, then delta alternating with periods of low-voltage flattening or fast activity (Schneck, 1965).

As the condition advances, initially multifocal then bilaterally synchronous epileptiform potentials become increasingly frequent. These may be in the form of spike–wave complexes but more commonly resemble SSWs and bisynchronous spikes. The background is slow and arrhythmic.

Neuronal Ceroid Lipofuscinosis

As with other degenerative conditions, the alpha activity slows and then gradually disappears. Rolandic activity also is disrupted, but less so than the alpha (Pampiglione & Harden, 1973). Bursts and runs of 2 to 7 Hz waves attaining 200 to 500 µV dominate the tracing. Early in the course, passive eye closure augments delta activity posteriorly; later, there is no effect. Diffuse bisynchronous epileptiform paroxysms in the form of spike–waves, SSWs, spikes, or sharp waves are mingled with the slow waves.

Pampiglione and Harden (1973) described bilaterally synchronous, high-voltage (50 to 500 µV) spike discharges with myoclonus in response to photic flashes at rates of 3 per second or less. Amplitude of the discharges progressively diminishes at flash rates above 4 per second.

Background slowing, diffuse delta, and bisynchronous epileptiform paroxysms all increase in prominence as the disease advances. However, as the terminal phase approaches, the amplitude of activity declines; paroxysmal activity disappears despite the persistence of clinical myoclonus.

Familial Myoclonus Epilepsy (Unverricht's Myoclonus Epilepsy)

Loss of normal background rhythms, prominent generalized spike–waves and polyspike–waves, and diffuse delta also are features of familial myoclonus epilepsy. Such generalized epileptiform paroxysms and their associated diffuse myoclonus are easily activated by afferent stimuli, especially photic stimulation (Genton & Roger, 1993). Any flash rate may provoke these discharges.

Lafora Disease

Slower waves at 3 to 6 per second become superimposed upon a progressively disorganized background. Bisynchronous spike–waves and polyspike–waves appear spontaneously in abundance. Photic stimulation readily elicits them as well. Myoclonus may or may not be temporally related to spikes. Visual seizures have been documented with unilateral occipital–posterior–temporal sequential spikes (Tassinari et al., 1978).

Rasmussen Encephalitis

EEGs reflect the aggressiveness of this process, with prominent and persistent arrhythmic delta waves, loss of background features such as alpha and mu rhythms, and the presence of abundant spikes. The latter may be periodic and may propagate widely throughout the involved hemisphere. Such propagation may explain the often complex ictal symptoms that implicate widely separate portions of the same hemisphere. For example, visual or somatosensory symptoms may proceed to tonic asymmetrical motor seizures. Scrutiny of awake and sleep recordings may reveal less prominent but independently occurring spikes or other abnormalities in


the healthier hemisphere, leading to the term “regionally accentuated encephalitis.” Bilaterally synchronous spike–waves may occur.

Acquired Epileptic Aphasia (Landau–Kleffner Syndrome)

Abundant spikes or spike–wave complexes appear bilaterally with predominance over the temporal, parietal, and occipital regions in acquired epileptic aphasia, with the emphasis shifting from side to side. Background activity is normal (Hirsch et al., 1990). Sleep onset and non-REM sleep may augment spike quantity, but they may persist in REM sleep (Roger et al., 1993). Such EEG abnormalities become less prominent in adolescents, which is roughly parallel with the decline in the disorder for most patients. This syndrome and electrographic status epilepticus of sleep may be variants of the same disorder.

Rett Syndrome

The EEG becomes progressively abnormal in this degenerative corencephalopathy. During stage 2, when 70% to 80% of these girls have one or more seizure types, central–parietal spikes appear; they are almost continuous in sleep. Hand movements block these spikes (Niedermeyer & Naidu, 1990). SSWs are a major feature; these waves usually achieve maximum expression posteriorly, as compared with the usual anterior field distribution of SSWs in the Lennox–Gastaut syndrome (Niedermeyer & Naidu, 1987). Multifocal spikes also may appear. Trauner and Haas (1985) found disorganized, slow background activity during wakefulness and quasiperiodic bursts of high-amplitude delta with interspersed epochs of attenuation lasting 3 to 4 s.

Subacute Sclerosing Panencephalitis: A Grey and White Matter Disease

Subacute sclerosing panencephalitis is a slowly progressive encephalitis caused by persistent infection with measles virus. It is characterized by mental deterioration, visual impairment, seizures, myoclonus, and movement disorders. The recent decline in its incidence may relate to prevention of measles through widespread use of measles virus vaccine.

Whenever the EEG of a young patient with any part of this syndrome contains high-voltage periodic complexes, further diagnostic efforts should be pursued. The distinguishing features of such complexes are their constant form in an individual patient on a single occasion, the regular repetition rate during a single recording, and their constant relationship to myoclonic jerks when they are present (Cobb, 1966). The complexes consist of 100 to 1000 µV, 1 to 3 Hz waves, sometimes with intermingled spikes or sharp waves. On different recordings, their duration ranges from 1 to 3 s and the intercomplex interval varies from 2 to 20 s. Longer intervals may be present early in the course of the disease (Reiher et al., 1973). Although the periodic complexes are usually diffuse and symmetrical, they may less commonly be either hemispheric or even regional.

Myoclonus, actually brief (<100 ms) myotonic events when present, almost always maintains a fixed temporal relationship to the complexes. However, the jerks may precede, follow, or occur simultaneously with the EEG events. Periodic complexes may appear without myoclonus, but the converse is very rare. Myoclonus usually ceases in deep sleep, whereas the complexes persist. Afferent stimuli of any kind usually do not disturb the rhythmicity of either the complexes or the myoclonus. In certain instances, the repetitive event accompanying the repetitive EEG complexes may be a sudden loss of muscle tone.

Early in the course of the illness, the periodic complexes may arise out of normal background activity. As the disease progresses, excessive theta and delta appear. However, the background deterioration is not steady; normal recordings may be interspersed among those with periodic complexes early in the course of the disease. The amplitude of the background activity declines later in the illness when the complexes disappear.

Uncommonly, spike–wave complexes and SSWs may appear, but they usually remain independent of the periodic EEG phenomena (Cobb, 1966; Westmoreland et al., 1976).

The sleep cycle may simplify to a low-voltage fast pattern and a high-voltage delta phase. Sleep spindles decline near the terminal part of the illness (Petre-Quadens et al., 1968). Periodic complexes may appear only in sleep during certain epochs in some patients (Westmoreland et al., 1977).

Brain Tumors

The classic EEG sign of brain tumor is persistent regional delta activity with spike discharges in the same region if the tumor is growing slowly. Thus, we found persistent EEG delta activity in 10 of 16 patients whose tumors presented as chronic uncontrolled partial seizure disorders (Blume et al., 1982). However, multiple independent spike discharges occurred in a majority of epileptic patients with tumors; 4 of the 16 with tumor had generalized spike–wave discharges. Thus, the type and distribution of epileptiform discharges does not distinguish patients with tumors but persistent focal delta activity over several recordings may suggest their presence. Of course, improved neuroimaging has lessened the EEG's value in tumor detection.




Aicardi J, Chevrie JJ. Myoclonic epilepsies of childhood. Neuropaediatrie. 1971; 3:177–190.

Aicardi J, Chevrie JJ. The significance of electroencephalographic paroxysms in children less than 3 years of age. Epilepsia. 1973;14:47–55.

Alvarez N, Medina C, Lombroso CT. Paroxysmal spike and slow wave bursts of early sleep in children: A controlled study. Electroencephalogr Clin Neurophysiol. 1980;49:23(P)–24(P).

Ashwall S, Schneider S. Failure of electroencephalography to diagnose brain death in comatose children. Arch Neurol. 1979;6:512–517.

Bauer G. Coma and brain death. In: Niedermeyer E, Lopes da Silva F, eds. Electroencephalography: Basic Principles, Clinical Applications, and Related Fields. Baltimore: Urban & Schwarzenberg; 1987:391–404.

Berkovic SF, Andermann F, Carpenter S, Wolfe LS. Progressive myoclonus epilepsies: Specific causes and diagnosis. N Engl J Med. 1986;315:296–305.

Bickford RG, Butt HR. Hepatic coma: The electroencephalographic pattern. J Clin Invest. 1955;34:790–799.

Blume WT. Abnormal EEG: Epileptiform potentials. In: Blume WT. Atlas of Pediatric Electroencephalography. New York: Raven Press, 1982;141–142.

Blume WT, Girvin JP, Kaufmann JCE. Childhood brain tumors presenting as chronic and controlled focal seizure disorders. Ann Neurol. 1982;12:538–541.

Blume WT, Girvin JP, McLachlan RS, Gilmore BE. Effective temporal lobectomy in childhood without invasive EEG. Epilepsia. 1997;38:164–167.

Blume WT, Kaibara M. Localization of epileptic foci in children. Can J Neurol Sci. 1991;18:570–572.

Blume WT, McNeill DK, Young GB, Chiu P. The low-voltage EEG in coma. In: Machado C, ed. Brain Death. Amsterdam: Elsevier; 1995:157–161.

Bobele GB, Steele K, Leonard J. Determination of brain death in newborns and infants: A comparison of electroencephalograms and radionuclide cerebral perfusion scans. Ann Neurol. 1993;34:478.

Boyd SG, Harden A, Patton MA. The EEG in early diagnosis of the Angelman (happy puppet) syndrome. Eur J Pediatr. 1988;147:508–513.

Braathen G, Melander H. Early discontinuation of treatment in children with uncomplicated epilepsy: A prospective study with a model for prediction of outcome. Epilepsia. 1997;38:561–569.

Camfield PR, Camfield CS, Dooley JM, Tibbles JA, Fung T, Garner B. Epilepsy after a first unprovoked seizure in childhood. Neurology. 1985;35:1657–1660.

Camfield PR, Metrakos K, Andermann E. Basilar migraine, seizures, and severe epileptiform EEG abnormalities. Neurology. 1978;28:584–588.

Carpay JA, de Weerd AW, Schimsheimer RJ, et al. The diagnostic yield of a second EEG after partial sleep deprivation: A prospective study in children with newly diagnosed seizures. Epilepsia. 1997;38:595–599.

Chatrian GE, Shaw CM, Leffman H. The significance of periodic lateralized epileptiform discharges in EEG: An electrographic, clinical and pathological study. Electroencephalogr Clin Neurophysiol. 1964;17:177–193.

Cobb W. The periodic events of subacute sclerosing leukoencephalitis. Electroencephalogr Clin Neurophysiol. 1966;21:278–294.

Dinner DS, Luders H, Rothner AD, Erenberg G. Complex partial seizures of childhood onset: A clinical and encephalographic study. Cleve Clin. 1984;51:287–291.

Donat JF, Wright FS. Episodic symptoms mistaken for seizures in the neurologically impaired child. Neurology. 1990;40:156–157.

Dorenbaum D, Mencel E, Blume WT, Fisman S. EEG findings and language patterns in autistic children: Clinical correlations. Can J Psychiatry. 1987;32:31–34.

Frantzen E, Harvald B, Haugsted H. Fresh head injuries. Acta Psychiatr Neurol Scand. 1958;33:417–428.

Frantzen E, Lennox-Buchthal M, Nygaard A. Longitudinal EEG and clinical study of children with febrile convulsions. Electroencephalogr Clin Neurophysiol. 1968; 24:197–212.

Frantzen E, Lennox-Buchthal M, Nygaard A, Stene J. A genetic study of febrile convulsions. Neurology. 1970;20:909–917.

Froelich WA, Carter CC, O'Leary JL, Rosenbaum HE. Headache in childhood. Electroencephalographic evaluation of 500 cases. Neurology. 1960;10:639–642.

Frost JD Jr, Hrachovy RA, Glaze DG, McCully MI. Sleep modulation of interictal spike configuration in untreated children with partial seizures. Epilepsia. 1991;32:341–346.

Gastaut H, Broughton R. Epileptic seizures: Clinical and electrographic features and pathophysiology. Springfield, IL: Charles C. Thomas; 1972:41–45.

Gastaut H, Poirier F, Payan H, Salomon G, Toga M, Vigouroux M. HHE syndrome: Hemiconvulsions, hemiplegia, epilepsy. Epilepsia. 1960;1:418–447.

Gastaut H, Zifkin BG. Benign epilepsy of childhood with occipital spike and wave complexes. In: Andermann F, Lugaresi E, eds. Migraine and Epilepsy. Boston: Butterworth–Heinemann; 1987:47–81.

Genten P, Roger J. The progressive myoclonus epilepsies. In: Wyllie E ed, The Treatment of Epilepsy: Principles and Practice. Philadelphia: Lea & Febiger, 1993: 571–583.

Gibbs FA, Gibbs EL. Atlas of Electroencephalography. Vol 3: Neurological and Psychiatric Disorders. Reading: Addison–Wesley; 1964:185–198.

Gloor P, Kalabay O, Giard N. The electroencephalogram in diffuse encephalopathies: Electroencephalographic correlates of grey and white matter lesions. Brain. 1968;91:779–802.

Hirsch E, Marescaux C, Maquet P, et al. Landau–Kleffner syndrome: A clinical and EEG study of five cases. Epilepsia. 1990;31:756–767.

Jorgenson EO. Brain death: Retrospective surveys. Lancet. 1981;1:378–379.

Kellaway P. Head injury in children. Electroencephalogr Clin Neurophysiol. 1955;7: 497(P)–498(P).

Kooi KA, Tucker RP, Marshall RE. Fundamentals of electroencephalography, 2nd ed. Hagerstown, MD: Harper & Row; 1978:155–167.



Laan LA, Renier WO, Arts WF, et al. Evolution of epilepsy and EEG findings in Angelman syndrome. Epilepsia. 1997;38:195–199.

Laplane R, Salbreux R. Les convulsions hyperpyretiques. Rev Prat. 1963;13: 753–761.

Metrick ME, Ritter FJ, Gates JR, Jacobs MP, Skare SS, Loewenson RB. Nonepileptic events in childhood. Epilepsia. 1991;32:322–328.

Miller H, Blume WT. Primary generalized seizure disorder: correlation of epileptiform discharges with seizure frequency. Epilepsia. 1993;34:128–132.

Mohandas A, Chou SN. Brain death: A clinical and pathological study. J Neurosurg. 1971;35:211–218.

Moshe SL, Alvarez LA. Diagnosis of brain death in children. J Clin Neurophysiol. 1986;3:239–249.

Niedermeyer E, Naidu S. Degenerative disorders of the central nervous system. In: Niedermeyer E, Lopes da Silva F, eds. Electroencephalography: Basic Principles, Clinical Applications and Related Fields. Baltimore: Urban & Schwarzenberg; 1987:317–338.

Niedermeyer E, Naidu S. Further EEG observations in children with the Rett syndrome. Brain Dev. 1990;12:53–54.

Otsubo H, Steinlin M, Hwang PA, et al. Positive epileptiform discharges in children with neuronal migration disorders. Pediatr Neurol. 1997;16:23–31.

Pallis C. ABC of brain death: The position in the USA and elsewhere. Br Med J. 1983;286:209–210.

Palmini A, Andermann F, Olivier A, et al. Focal neuronal migration disorders and intractable partial epilepsy: A study of 30 patients. Ann Neurol. 1991a;30:741–749.

Palmini A, Andermann F, Olivier A, Tampieri D, Robitaille Y. Focal neuronal migration disorders and intractable partial epilepsy: Results of surgical treatment. Ann Neurol. 1991b;30:750–757.

Pampiglione G, Harden A. Resuscitation after cardiocirculatory arrest. Lancet. 1968;1:1261–1265.

Pampiglione G, Harden A. Neurophysiological identification of a late infantile form of “neuronal lipidosis.” J Neurol Neurosurg Psychiatry. 1973;36:68–74.

Pathak P, Blume WT. The asymmetrical epileptogenicity of brain lesions. Can J Neurol Sci. 1997;24(Suppl 1):515–516(absract).

Petre-Quadens O, Sfaello Z, van Bogaert L, Moya G. Sleep study in SSPE (first results). Neurology. 1968;18:60–68.

Prensky AL, Sommer D. Diagnosis and treatment of migraine in children. Neurology. 1979;29:506–510.

Rasmussen T. Surgery for epilepsy arising in regions other than the temporal and frontal lobes. In: Purpura DP, Penry JK, Walter RD, eds. Advances in Neurology. Vol 8:Neurosurgical Management of the Epilepsies. New York: Raven Press; 1975:207–226.

Raymond AA, Fish DR, Boyd SO, Smith SJ, Pitt MC, Kendall B. Cortical dysgenesis: Serial EEG findings in children and adults. Electroencephalogr Clin Neurophysiol. 1995;94:389–397.

Reiher J, Lapointe LR, Lessard L. Prolonged and variable intervals between EEG complexes in subacute inclusion body encephalitis. Can Med Assoc J. 1973; 108:729–732.

Robinson RO. Brain death in children. Arch Dis Child. 1981;56:657–658.

Roger J, Genton P, Bureau M, Dravet C. Less common epileptic syndrome. In: Wyllie E, ed. The Treatment of Epilepsy: Principles and Practice. Philadelphia: Lea & Febiger; 1993:624–635.

Rossi PG, Parmeggiani A, Bach V, Santucci M, Visconti P. EEG features and epilepsy in patients with autism [see comments]. Brain Dev. 1995;17:169–174.

Rowland TW, Donnelly JH, Jackson AH, Jamroz SB. Brain death in the pediatric intensive care unit. A clinical definition. Am J Dis Child. 1983;137:547–550.

Saunders MG, Westmoreland BE. The EEG in evaluation of disorders affecting the brain diffusely. In: Klass DW, Daly DD, eds. Current Practice of Clinical Electroencephalography. New York: Raven Press; 1979:370–371.

Schneck L. The early electroencephalographic and seizure characteristics of Tay–Sachs disease. Acta Neurol Scand. 1965;41:163–171.

Seshia SS, Chow PN, Sankaran K. Coma following cardiorespiratory arrest in childhood. Dev Med Child Neurol. 1979;21:143–153.

Shinnar S, Berg AT, Moshe SL, et al. Risk of seizure recurrence following a first unprovoked seizure in childhood: A prospective study. Pediatrics. 1990;85: 1076–1085.

Silverman D. Electroencephalographic study of acute head injury in children. Neurology. 1962;12:273–281.

Smith JB, Westmoreland BE, Reagen TJ, Sandok BA. A distinctive clinical EEG profile in herpes simplex encephalitis. Mayo Clin Proc. 1975;50:469–474.

Sundaram MBM, Blume WT. Triphasic waves: Clinical correlates and morphology. Can J Neurol Sci. 1987;14:136–140.

Tassinari CA, Bureau-Paillas M, Dalla Bernardina B, et al. La maladie de Lafora. Rev EEG Neurophysiol. 1978;8:107–122.

Tomlin PJ, Martin JW, Honigsberger L. Brain death: retrospective surveys. Lancet. 1981;1:378.

Trauner DA, Haas RH. Electroencephalographic abnormalities in Rett's syndrome. Ann Neurol. 1985;18:394.

Upton A, Gumpert J. Electroencephalography in diagnosis of herpes simplex encephalitis. Lancet. 1970;1:650–652.

Westmoreland BE, Blume WT, Gomez MR. Generalized sharp and slow wave and electrodecremental seizure pattern in subacute sclerosing panencephalitis. Mayo Clin Proc. 1976;51:107–111.

Westmoreland BE, Gomez MR, Blume WT. Activation of periodic complexes of subacute sclerosing panencephalitis by sleep. Ann Neurol. 1977;1: 185–187.

Young GB. Major syndromes of impaired consciousness. In: Young GB, Ropper AH, Bolton CF, eds. Coma and Impaired Consciousness: A Clinical Perspective. New York: McGraw-Hill; 1998:39–78.