Neonatal epileptic seizures occur from birth to the end of the neonatal period.This is the most vulnerable of all the other periods of life for the development of epileptic seizures, particularly in the first 1 or 2 days from birth. Neonatal seizures differ from those of older children and adults. They may be short-lived events lasting for just a few days, but they often signify serious malfunction or damage of the immature brain, and constitute a neurological emergency that demands urgent diagnosis and management. Most neonatal seizures are acute (pro-voked, occasional, reactive) symptomatic seizures caused by an acute illness such as hypoxic–ischaemic encephalopathy, stroke or infection. Seizures are the most common and important sign of acute neonatal encephalopathy; they are a major risk for death or subsequent neurological disability and, by themselves, may contribute to an adverse neuro-developmental outcome.
- 1. Clarifications on classification and terminology
- 2. Demographic data
- 3. Clinical manifestations
- 4. Aetiology
- 5. Diagnostic procedures
- 6. Differential diagnosis
- 7. Prognosis
- 8. Management
- 9. References
Clarifications on classification and terminology
The ILAE Commission (1989) broadly classifies the neonatal seizures among ‘epilepsies and syndromes undetermined as to whether they are focal or generalised’ under the subheading ‘with both generalised and focal seizures’. The new ILAE report stated the following: Although the components of neonatal seizures can be described in terms of the seizure types itemized in the list of epileptic seizures , they often display unique organizational features. Therefore, a study group will be created to more completely define and characterize the various types of neonatal seizures (Engel, 2006). In the ILAE (2010) revised classification neonatal seizures are no longer regarded as a separate entity. Seizures in neonates can be classified within the proposed scheme.
- The neonatal period is defined as the first 28 days of life of a full-term infant.
- Neonatal seizures are those that occur from birth to the end of the neonatal period.
- Gestrational age is defined as the duration of pregnancy.
- Chronological age is the actual ‘legal’ age of the infant from time of birth.
- Conceptional age is the combined destational and chronological ages.
- Full-term infants are those with 40 weeks’ gestational age.
The prevalence of neonatal seizures is approximately 1.5% and overall incidence approximately 3/1000 live births. The incidence in pre-term infants is very high (57–132/1000 live births). Of neonatal seizures 80% occur in the first 1 or 2 days during the first week of life.
Neonatal seizures are paroxysmal, repetitive and stereotypical events. They are usually clinically subtle, inconspicuous and difficult to recognise from the normal behaviours of the inter-ictal periods or physiological phenomena. There is no recognisable post-ictal state. Generalised tonic–clonic seizures (GTCSs) are exceptional or may not occur (Korff & Nordli, 2005).
The most widely used scheme is by Volpe (Volpe, 1989) and constitutes five main types of neonatal seizure:
- subtle seizures (50%)
- tonic seizures (5%)
- clonic seizures (25%)
- myoclonic seizures (20%)
- non-paroxysmal repetitive behaviours.
- focal clonic
- focal tonic
- generalised tonic
- motor automatisms (which include ocular signs, and oral–buccal–lingual, progression and complex purposeless movements).
Almost a quarter of infants experience several seizure types and the same seizure may manifest with subtle, clonic, myoclonic, autonomic or other symptoms (See Fig.1).
Figure 1 (A,B) Apnoeic, myoclonic, clonic and subtle seizure of motor automatisms associated with various ictal EEG patterns and locations. (C) ‘Electroclinical dissociation’: the electrical discharge is not associated with apparent clinical manifestations.
Scher (Scher, 2006) has proposed a multi-dimensional classification scheme for neonatal seizures that will help strategise specific therapeutic interventions in order to optimise neurological outcome and anticipate later neurological morbidities, including epilepsy risk. This scheme combines epileptic and non-epileptic seizure descriptions that capture time-specific and brain region-specific mechanisms for seizures.
Subtle seizures are far more common than other types of neonatal seizures. They are described as subtle because the clinical manifestations are frequently overlooked. They imitate normal behaviours and reactions. Subtle seizures manifest with the following:
- ocular movements, which range from random and roving eye movements to sustained conjugate tonic deviation with or without jerking. Eyelid blinking or fluttering, eyes rolling up, eye opening, fixation of a gaze or nystagmus may occur alone or with other ictal manifestations
- oral–buccal–lingual movements (sucking, chewing, smacking and tongue protrusions)
- progression movements (rowing, swimming, pedalling, bicycling, thrashing or struggling movements)
- complex purposeless movements (sudden arousal with episodic limb hyperactivity and crying) (Mizrahi & Watanabe, 2005).
Clonic seizures are rhythmic jerks that may localise in a small part of the face or limbs, axial muscles and the diaphragm, or be multifocal or hemiconvulsive. Multifocal clonic seizures may migrate to other body parts or other limbs. Todd’s paresis follows prolonged hemiconvulsions.
Tonic seizures manifest with sustained contraction of facial, limb, axial and other muscles. They may be focal, multifocal or generalised, symmetrical or asym-metrical. Truncal or limb tonic extensions imitate decerebrate or decorticate posturing. These occur particularly in pre-term infants and have a poor prognosis because they frequently accompany intra-ventricular haemorrhage.
Myoclonic seizures are rapid, single or arrhythmic repetitive jerks. They may affect a finger, a limb or the whole body. They may mimic Moro reflex and startling responses. They are more frequently in pre-term than in full-term infants indicating, if massive, major brain injury and poor prognosis(Watanabe et al., 1982). However, healthy pre-term and, although rarely, full-term neonates may have abundant myoclonic movements during sleep. Neonates have cortical, reticular and segmental types of myoclonus, similar to adult forms (Scher, 1985).
Spasms producing flexion or extension similar to those of West syndrome are rare. They are slower than myoclonic and clonic seizures and faster than tonic seizures (Fig.2).
Autonomic ictal manifestations
Autonomic ictal manifestations commonly occur with motor manifestations in 37% of subtle seizures. These are paroxysmal changes of heart rate, respiration and systemic blood pressure. Apnoea, as an isolated seizure phenomenon unaccompanied by other clinical epileptic features, is probably exceptional (Watanabe et al., 1982) Salivation and pupillary changes are common.
Duration of neonatal seizures
The duration of neonatal seizures is usually brief (10 s to 1–2 min) and repetitive with a median of 8 min in between each seizure. Longer seizures and status epilepticus develop more readily at this age, but convulsive neonatal status epilepticus is not as severe as that of older infants and children.
Non-epileptic neonatal seizures
Kellaway and Mizrahi (Mizrahi & Kellaway, 1999; Mizrahi, 2008) proposed that many of the subtle seizures, generalised tonic posturing and some myoclonic symptoms may be non-epileptic seizures. These show clinical similarities to reflex behaviours of the neonates, but are not associated with ictal EEG discharges and commonly correlate with diffuse abnormal brain processes such as hypoxicischaemic encephalopathy and a poor short-term outcome. They are considered as exaggerated reflex behaviours due to abnormal release of brain-stem tonic mechanisms from cortical control – hence, the term ‘brain-stem release phenomena’:
They most typically occur in neonates with clinical and EEG evidence of forebrain depression that may release brain stem facilitatory centres for generating reflex behaviours without cortical inhibition (Mizrahi & Kellaway, 1999; Mizrahi, 2008).
The aetiology of neonatal epileptic seizures is extensive and diverse (Table 1). Severe causes predominate. The prevalence and significance of aetiological factors are continually changing and are uneven between developed and resource-poor countries, depending on available improved neonatal and obstetric care. In most cases, the neonate may present with a combination of different neurological disturbances, each of which can cause seizures.
|Main causes of neonatal seizures|
|Haemorrhage and intracerebral infarction||++++|
|Malformations of cerebral development||+++|
|Drug withdrawal and toxic||+++|
|Inadvertent injections of local anaesthetics during delivery||+|
|Idiopathic benign neonatal seizures (familial and non-familial)||+|
Hypoxic–ischaemic encephalopathy is by far the most common cause – probably 80% of neonatal seizures.
Brain damage due to prenatal distress and malformations of cortical development is being recognised more frequently.
Intracranial haemorrhage and infarction, stroke and prenatal and neonatal infections are common.
Acute metabolic disturbances such as electrolyte and glucose abnormalities have been minimised because of improved neonatal intensive care and awareness of nutritional hazards. Late hypocalcaemia is virtually eliminated, whereas electrolytic derangement and hypoglycaemia are now rare.
Inborn errors of metabolism such as urea cycle disorders are rare.
Pyridoxine dependency, with seizures in the first days of life (reversible with treatment), is exceptional.
Exogenous causes of neonatal convulsions may be iatrogenic or due to drug withdrawal in babies born to mothers on drugs.
The increased risk to seizures of neonates may be due to a combination of factors specific to the developing brain that enhance excitation and diminish inhibition. There is an unequal distribution of anticonvulsant and proconvulsant neurotransmitters and networks (Velisek & Moshe, 2008; Holmes et al., 2002) (see also individual neonatal epileptic syndromes).
Neonatal seizures represent one of the very few emergencies in the newborn. Abnormal, repetitive and stereotypical behaviours of neonates should be suspected and evaluated as possible seizures. Polygraphic video-EEG recording of suspected events is probably mandatory for an incontrovertible seizure diagnosis. Confirmation of neonatal seizures should initiate urgent and appropriate clinical and laboratory evaluation for the aetiological cause (Table 1) and treatment. Family and pre-natal history is important. A thorough physical examination of the neonate should be coupled with urgent and comprehensive biochemical tests for correctable metabolic disturbances. Although rare, more severe inborn errors of metabolism should be considered for diagnosis and treatment.
Cranial ultrasonography and brain imaging with CT and preferably MRI (Johnston et al., 2008) should be used for the detection of structural abnormalities, such as malformations of cortical development, intracranial haemorrhage, hydrocephalus and cerebral infarction.
Cranial ultrasonography is the main imaging modality of mainly pre-term neonates. It is limited in resolution and the type of lesions that it can identify.
CT brain scans are now of high resolution and can be generated within seconds. They can accurately detect haemorrhage, infarction, gross malformations, and ven tricular and other pathological conditions. Sensitivity is low in malformations of cortical development.
MRI is the superior modality. MRI interpretation should take into consideration the normal developmental and maturational states of neonates and infants. In infants younger than 6 months, cortical abnormalities are detected with T2-weighted images, whereas T1-weighted ones are needed for the evaluation of brain maturation (Johnston et al., 2008)
The neonatal EEG is the most useful EEG application. Well-trained technologists and physicians are required. Polygraphic studies with simultaneous video-EEG recording are essential (Murray et al., 2008; Shany et al., 2006; Karayiannis et al., 2006).
Only 10% of neonates suspected of having seizures have EEG confirmation: clonic movements have the highest yield of 44%, but only 17% for ‘subtle’ movements.
Inter-ictal EEG epileptogenic spikes or sharp–slow-wave foci are not reliable markers in this age but certain inter-ictal EEG patterns may have diagnostic significance . These are:
- electrocerebral inactivity of a flat or almost-flat EEG of severe brain damage
- the burst-suppression pattern of neonatal epileptic encephalopathies
- theta pointu alternant of benign neonatal convulsions (Mizrahi & Tharp, 1982)
- persistently focal sharp or slow waves in localised lesions
- quasi-periodic focal or multifocal pattern in neonatal herpes simplex encephalitis (Mizrahi & Tharp, 1982)
- periodic complexes in glycine encephalopathy (Arzimanoglou et al., 2004)
- inter-hemispheric or intra-hemispheric abnormalities.
Background EEG activity, mainly in serial EEGs, often provides objective evidence of the degree and severity of the underlying cause.
The burst-suppression pattern is relatively frequent in the neonatal period. It is associated with heterogeneous seizures and can be inclued by drugs (Lombroso, 1990; Martínez-Bermejo et al., 2001) It is common in neonatal ischaemic encephalopathy, where it is usually transient and short-lived (‘trace paroxystique’ for French neonatologist) (Lombroso, 1990). Conversely, it is relatively stable, lasting for more than 2 weeks in Ohtahara syndrome and early myoclonic encephalopathy (Aicardi & Ohtahara, 2005).
Ictal EEG patterns vary significantly even in the same neonate and the same EEG recording (Figure 8.1). Ictal EEG paroxysms consist of repetitive waves with a predominant beta, alpha, theta and delta range, or a mixture of all, which may accelerate or decelerate in speed or both (Figures 1 and 3). They consist of spikes, and sharp, saw-tooth or sinusoidal waves (monomorphic or polymorphic), ranging in amplitude from very low to very high. The patterns may be synchronous or asynchronous, focal or multifocal and, less frequently, generalised. They may appear and disappear suddenly or build up from accelerating localised repetitive waves. Ictal discharges may gradually or abruptly change in frequency, amplitude and morphology in the course of the same or subsequent seizures. Conversely, they may remain virtually unchanged from onset to termination. The background EEG may be normal or abnormal.
Focal EEG ictal discharges usually associate with subtle, clonic or tonic seizures. The most common locations in order of prevalence are centrotemporal, occipital, midline (Cz) and temporal regions. Frontal localisations are exceptional. The same infant may have unifocal or multifocal ictal discharges, which may be simultaneous, develop one from another or occur independently in different brain sites. Clinical or EEG jacksonian march is not seen. Consistently focal EEG paroxysms are highly correlated with focal brain lesions. Seizures that lack or have an inconstant relationship with EEG discharges correlate with diffuse pathological conditions.
‘Zips’ is a descriptive term coined by Dr Panayitopoulos for a commonictal EEG pattern in neonates, which consists of localised episodic rapid spikes of accelerating and decelerating speed that look like zips (Figure 3). Zips may be associated with subtle and focal clonic and tonic seizures or remain clinically silent. Zips of subtle seizures are often multifocal and of shifting localisation.
Generalised ictal discharges are more likely to occur with myoclonic jerks and neonatal spasms.
Electroclinical dissociation or decoupling response
Only a fifth (21%) of electrical ictal EEG patterns associate with distinctive clinical manifestations (electroclinical seizures). All others are occult, i.e. they are clinically silent or subclinical (electrical or electrographic seizures) (Clancy et al., 1988).
Electrographic or electrical seizures, namely EEG electrical seizure activity without apparent clinical manifestations, are more common after initiation of anti-epileptic drugs (AEDs). This is because AEDs may suppress the clinical manifestations of seizures but not the EEG ictal discharge. This phenomenon is named the ‘decoupling response or ‘electroclinical dissociation’(Biagioni et al., 1988). Electroclinical dissociation may arise from foci not consistently reflected in surface electrodes. Neonates with electrographic seizures do not differ from those with exclusively electroclinical seizures with regard to aetiology or outcome, although the background EEG is more abnormal in the electrographic group (Weiner et al., 1991).Movements of the limbs occur at a statistically significant higher rate during electroclinical seizures. Electrographic seizures, similar to electroclinical seizures, are also associated with disturbed cerebral metabolism (Boylan et al., 1999).
Figure 3 (A,B) Continuous recording in a neonate with severe brain hypoxia. Zip-like discharges consist of high-frequency rapid spikes of accelerating and decelerating speed. They start from various locations, terminating in one while continuing in another EEG derivation. (C) Amplification of zips with high sensitivity in T6–O2 derivation. Note that the zips look like, but are not, muscle artefacts, which may also occur when zips are associated with motor seizure manifestations.
Electrical seizure patterns of usually poor prognosis
Alpha seizure discharges are characterised by sustained and rhythmic activity of 12 Hz and 20–70 μV in the centrotemporal regions.
Electrical seizure activity of the depressed brain is of low voltage, long duration, highly localised on one side and with little tendency to spread.
Stimulus-evoked electrographic or electroclinical patterns elicited by tactile or painful stimulation, usually occur in pre-term neonates or neonates with significant diffuse or multifocal brain damage (Scher, 1997). Most cases die or have significant neurological handicaps.
Post-ictal EEG usually returns to the pre-ictal state immediately (Figure 1). Transient slowing or depression of EEG activity may occur after frequent or prolonged seizures.
Neonatal seizures often impose significant difficulties in their recognition and differentiation from normal or abnormal behaviours of the pre-term and full-term neonate.
As a rule, any suspicious repetitive and stereotypical events should be considered as possible seizures requiring confirmation by video-EEG recording.
Normal behaviours: Among normal behaviours neonates may stretch, exhibit spontaneous sucking movements and have random and non-specific movements of the limbs. Intense physiological myoclonus may occur during rapid eye movement (REM) sleep. Jitteriness or tremulousness of the extremities or facial muscles are frequent in normal or abnormal neonates.
Tremor has a symmetrical ‘to and fro’ motion, is faster than clonic seizures, mainly affects all four limbs and will stop when the limb is restrained or repositioned. Conversely, clonic seizures are mainly focal, usually have a rate of 3 or 4 Hz or slower, decelerate in the progress of the attack and are not interrupted by passive movements.
Abnormal behaviours: Among abnormal behaviours of neonates with CNS disorders are episodic and repetitive oral–buccal–lingual movements. These are often reproducible with tactile or other stimuli and are interrupted by restraint. Conversely, neonatal seizures persist despite restraint and they are rarely stimulus sensitive.
Non-epileptic movement disorders: Neonatal seizures should be differentiated from benign neonatal sleep myoclonus, hyperekplexia and other non-epileptic movement disorders.
Significant impairment of vital signs, which may be periodic, is mainly due to non-neurological causes. Changes in respiration, the heart rate and blood pressure rarely occur as sole manifestations of neonatal seizures.
Inborn errors of metabolism manifest with neonatal subtle seizures or abnormal movements that may not be genuine epileptic seizures. Their identity is often revealed by other associated significant symptoms, such as peculiar odours, protein intolerance, acidosis, alkalosis, lethargy or stupor. In most cases, pregnancy, labour and delivery are normal. Food intolerance may be the earliest indication of a systemic abnormality. If untreated, metabolic disorders commonly lead to lethargy, coma and death. In surviving infants weight loss, poor growth and failure to thrive are common.
This is cause-dependent because the main factor that determines outcome is the underlying cause and not the seizures themselves. Despite high mortality (about 15% in developed and 40% in resource-poor countries) and morbidity rates (about 30% in developed countries), half the neonates with seizures achieve a normal or near-normal state. A third of the survivors develop epilepsy (Garcias et al., 2004) Table 2 provides indicators of good, bad or intermediate prognosis.
|Indicators of prognosis|
|Indicators of bad prognosis|
|Indicators of good prognosis|
|Indicators of intermediate or guarded prognosis|
Management demands accurate aetiological diagnosis and treatment of the cause of the seizures. The principles of general medical management and cardiovascular and respiratory stabilisation should be early and appropriately applied. Cardiorespiratory symptoms may result from the underlying disease, seizures and anti-epileptic medication.
Neonatal seizures of metabolic disturbances need correction of the underlying cause and not anti-epileptic medication. A trial of pyridoxine may be justifiable.
The drug treatment of neonatal seizures is empirical with significant practice variations among physicians. First phenobarbital and then phenytoin are the most commonly used AEDs. Large loading doses are followed by a maintenance scheme for a variable period.
The severity of the seizures appears to be a stronger predictor of the success of treatment than the assigned AED. Mild seizures or seizures decreasing in severity before treatment are more likely to respond regardless of the treatment assignment.
Facts and requirements for the treatment of neonatal epileptic seizures
Neonatal seizures have a high prevalence and their responses to AEDs is likely to be different to that of other specified groups of patients. Yet, current treatment of neonatal seizures is entirely empirical. Neonatologist rely on their medical judgement and ‘trials by success and error’ with off-label uses of both the newer and older AEDs.
The necessity for authorities, including formal regulatory agents, is self-evident.
Phenobarbital and phenytoin are equally effective. If either drug is given alone, the seizures are controlled in less than half of neonates. An RCT of phenobarbital versus placebo in a homogeneous group of newborns at high risk of developing early, sub clinical, EEG-detected, neo-natal seizures has been designed (Clancy, 2006).This study is intended to ‘affirm or refute the common practice of phenobarbital as the first-line treatment of neonatal seizures’.
Fosphenytoin is an attractive alternative to phenytoin because of its lesser potential for adverse reactions at the infusion site and the facility for intramuscular administration.
Intravenous benzodiazepines such as diazepam, lorazepam, clonazepam and midazolam are used, particularly in Europe, for acute neonatal seizures, although in a recent controlled study the results with benzodiazepines as second-line treatment were not encouraging (Boylan et al., 2004).
Primidone, valproate, lidocaine, carbamazepine, and paraldehyde are also used mainly as adjunctive AEDs if others fail (paraldehyde is now no longer available in the USA).
Newer AEDs are not licensed in the treatment of neonatal seizures. However, a recent survey in the USA showed that 73% of paediatric neurologists recommended one or both of levetiracetam (47%) and topiramate (55%) and made different dosing recommendations. Respondents considered both agents to be efficacious in the majority of cases; adverse effects were recognized more frequently with topiramate.40 Lamotrigine is also used (Barr et al., 1999)
Maintenance treatment: This may not be needed or may be brief because the active seizure period in neonates is usually short. Less than 15% of infants with neonatal seizures will have recurrent seizures after the newborn period (Massingale & Buttross, 1993). A normal EEG and other predictors of good outcome may encourage early discontinuation of treatment. The current trend is to withdraw the AED 2 weeks after the last seizure.
Do electrographic (electrical) seizures need treatment?
There is a significant difference of opinion as to whether EEG electrical seizure activity that may persist despite drug control of clinical seizures needs more vigorous treatment. Electrical seizures may be highly resistant to drug treatment and attempts to eliminate them may require high doses of usually multiple drugs, with significant adverse reactions such as CNS or respiratory depression and systemic hypotension. The risks should be weighed against the benefits while also remembering that these will eventually subside.
(1989) Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League Against Epilepsy.Epilepsia 30 (4):389-99. PMID: 2502382.
Arzimanoglou A, Guerrini R, Aicardi J. Aicardi’s epilepsy in children. Philadelphia: Lippincott Williams & Wilkins, 2004.
Boylan GB, Panerai RB, Rennie JM, Evans DH, Rabe-Hesketh S, Binnie CD (1999) Cerebral blood flow velocity during neonatal seizures. Arch Dis Child Fetal Neonatal Ed 80 (2):F105-10. PMID: 10325785.
Boylan GB, Rennie JM, Chorley G, Pressler RM, Fox GF, Farrer K et al. (2004) Second-line anticonvulsant treatment of neonatal seizures: a video-EEG monitoring study. Neurology 62 (3):486-8. PMID: 14872039.
Barr PA, Buettiker VE, Antony JH (1999) Efficacy of lamotrigine in refractory neonatal seizures. Pediatr Neurol 20 (2):161-3. PMID: 10082350.
Garcias Da Silva LF, Nunes ML, Da Costa JC (2004) Risk factors for developing epilepsy after neonatal seizures. Pediatr Neurol 30 (4):271-7. DOI: 10.1016/j.pediatrneurol.2003.09.015 PMID: 15087106.
Johnston JM, Smyth MD, McKinstry RC. Basics of neuroimaging in pediatric epilepsy. In: Pellock JM, Bourgeois BFD, Dodson WE, eds. Pediatric epilepsy: Diagnosis and treatment (third edition), pp 213-26. New York: Demos Medical Publishing, 2008.
Karayiannis NB, Tao G, Frost JD, Jr., Wise MS, Hrachovy RA, Mizrahi EM. Automated detection of videotaped neonatal seizures based on motion segmentation methods. Clin Neurophysiol 2006;117:1585–94.
Korff C, Nordli DR (2005) Do generalized tonic-clonic seizures in infancy exist?Neurology 65 (11):1750-3. DOI: 10.1212/01.wnl.0000187125.87414.f3 PMID: 16344517.
Lombroso CT. Early myoclonic encephalopathy, early infantile epileptic encephalopathy, and benign and severe infantile myoclonic epilepsies: a critical review and personal Neuropsychol 1990;7:380–408.
Massingale TW, Buttross S (1993) Survey of treatment practices for neonatal seizures. J Perinatol 13 (2):107-10. PMID: 8515301.
Murray DM, Boylan GB, Ali I, Ryan CA, Murphy BP, Connolly S. Defining the gap between electrographic seizure burden, clinical expression and staff recognition of neonatal seizures. Arch Dis Child Fetal Neonatal Ed 2008;93:F187–F191
Mizrahi EM, Kellaway P. Diagnosis and management of neonatal seizures. Hagerstown: Lippincott Williams & Wilkins, 1999.
Mizrahi EM. Neonatal seizures. In: Pellock JM, Bourgeois BFD, Dodson WE, eds. Pediatric epilepsy:Diagnosis and treatment (third edition) pp. 229-40. New York: Demos Medical Publishing, 2008.
Mizrahi EM, Watanabe K. Symptomatic neonatal seizures. In: Roger J, Bureau M, Dravet C, Genton P, Tassinari CA, Wolf P, eds. Epileptic syndromes in infancy, childhood and adolescence. Fourth edition, pp 17–38. Montrouge, John Libbey & Co. Ltd, 2005.
Mizrahi EM, Tharp BR (1982) A characteristic EEG pattern in neonatal herpes simplex encephalitis.Neurology 32 (11):1215-20. PMID: 6890154.
Martínez-Bermejo A, Roche C, López-Martín V, Arcas J, Tendero A (2001) [Neonatal EEG trace of burst suppression. Etiological and evolutionary factors]. Rev Neurol 33 (6):514-8. PMID: 11727229.
Scher MS (1985) Pathologic myoclonus of the newborn: electrographic and clinical correlations.Pediatr Neurol 1 (6):342-8. PMID: 3939749.
Scher MS. Neonatal seizure classification: A fetal perspective concerning childhood epilepsy. Epilepsy Res 2006;70 Suppl:41–57.
Shany E, Khvatskin S, Golan A, Karplus M. Amplitude-integrated electroencephalography: a tool for monitoring neonates. Pediatr Neurol 2006;34:194–9.
Scher MS (1997) Stimulus-evoked electrographic patterns in neonates: an abnormal form of reactivity.Electroencephalogr Clin Neurophysiol 103 (6):679-91. PMID: 9546495.
Velisek L, Moshe SL. Pathophysiology of seizures and epilepsy in the immature brain: cells, synapses and circuits. In: Pellock JM, Dodson WE, Bourgeois BFD, eds. Pediatric epilepsy: Diagnosis and treatment (third edition), pp 3–30. New York: Demos Medical Publishing, 2008.
Volpe JJ (1989) Neonatal seizures: current concepts and revised classification.Pediatrics 84 (3):422-8. PMID: 2671912.