The neonatal period is conventionally limited to the first 28 days of life. However, some seizures with onset in the first 2 to 3 months of life display similar features and will be considered here. The vast majority of neonatal seizures, however, occur during the first 10 days of life. Most neonatal seizures are short-lived events, lasting only a few days, and usually do not herald a chronic convulsive disorder unless structural damage is present whether due to malformations or acquired disease. Therefore, the term epilepsy, meaning a chronic, recurrent disorder with unprovoked paroxysmal events, is only rarely adequate, and the noncommittal term of neonatal seizures or convulsions is more appropriate. Seizures are one of the most common neurologic problems in the first days of life. Their real frequency is not known because it is becoming clear that all abnormal, stereotyped, and periodically recurring movements may not be epileptic seizures in the sense of excessive discharges of the gray matter, but they can be due to other mechanisms, such as release of brainstem tonic mechanisms from cortical control as a result of lesions or dysfunction of the cortex.⁵ The clinical features of neonatal seizures are often relatively unimpressive (see Chapter 56). Generalized seizures are rare, although occasionally cases of infantile spasms or massive myoclonias may occur.^1,3 Focal or multifocal seizures are the rule and several types are recognized. Focal clonic seizures may remain in a fixed location or may involve several focal areas at the same time or in succession (migratory seizures). Clonic seizures are usually associated with typical rhythmic electroencephalographic (EEG) discharges, although these may also occur with tonic or subtle seizures. Tonic seizures may be localized to a single segment; when generalized, they are more often the result of nonepileptic tonic liberation than truly epileptic. Subtle seizures,³⁰ sometimes referred to as motor automatisms,¹⁹ are frequent and may include random and roving eye movements, sucking, chewing motions, tongue protrusion, rowing or swimming or boxing movements of the arms, and pedaling or bicycling movement of the lower limbs. Apneic seizures are relatively common.³¹ Although some subtle seizures are associated with rhythmic ictal EEG discharges and are clearly epileptic, ictal EEG often does not show typical epileptic activity, and the nature of the clinical events is difficult to determine. In some cases, such ictal phenomena can be provoked by external stimuli that can demonstrate temporal and spatial summation, or they can be inhibited by restraint or repositioning of the involved part. Such features may be more suggestive of brainstem release phenomena, and these events are mainly interpreted as nonepileptic. The absence of an EEG discharge, however, does not completely rule out the possibility of an epileptic phenomenon, thus raising a difficult diagnostic problem.

Ictal EEG discharges in neonates may be highly polymorphic. The two main components, which may be associated withone another, are repetitive sharp waves or spikes and abnormal paroxysmal rhythms including beta, alpha, theta, or delta rhythms that usually remain focal or involve only one hemisphere.^13,16,21 Almost all paroxysmal EEG activity in the neonate begins focally. Ictal discharges are extremely variable in appearance, voltage, frequency, and polarity, and their aspect can change suddenly. Unrelated discharges of various shapes and rhythms frequently occur independently at different locations in both hemispheres. Bilateral symmetric discharges are rare and are associated specifically with unusual ictal phenomena such as myoclonias or spasms. In benign familial neonatal convulsions, they may consist of generalized flattening of the tracing, followed by high-amplitude, slow theta waves,^12,26 but these events follow a focal onset.

Although such differences between neonatal epileptic versus nonepileptic seizures can be of practical value because the use of anticonvulsant drugs that depress cortical activity might be contraindicated in nonepileptic release phenomena,⁵ a firm differentiation of epileptic from nonepileptic seizures in neonates may be impossible.

From a clinical viewpoint, seizures must be differentiated from jittering or tremulations, shuddering, and benign neonatal sleep myoclonus.

Ictal EEG recording is helpful when showing clear paroxysmal activity. Ictal EEG events may not be easily differentiated from other patterns that are not infrequently seen in the neonatal EEG. These include brief intermittent repetitive discharges (BIRDS). Most investigators, therefore, require discharges of at least 10 seconds to accept their ictal nature.^28,29 Isolated interictal sharp waves should not be regarded as pointing necessarily to a diagnosis of seizures. However, some seizures with all the clinical features of classical epileptic seizures are not accompanied by typical discharges on the scalp.⁸ Conversely, many characteristic EEG discharges are not associated with any clinical manifestation,¹³ and such electroclinical dissociation seems to occur especially following long series of seizures. It is not currently possible to decide whether such purely electrical seizures have the same significance as electroclinical attacks, carry the same risk of brain damage, and require the same treatment.

Neonatal seizures rarely occur as isolated events. In most cases, the seizures are frequently repeated over a period of a few days and may result in status epilepticus.²⁸ They then subside, irrespective of whether the cause is a major brain lesion—as in hypoxic-ischemic encephalopathy—or a benign, purely functional disturbance—as in benign familial neonatal convulsions. The duration of individual seizures is usually brief, even in cases of status. An average duration of 137 ± 11 sec was found in one study.⁷ This brief duration may be one reason for the absence of residual damage in such cases as hypocalcemic seizures that leave no residua even if they are repeated over several days.

The causes of neonatal seizures differ considerably from those of seizures in older children. A majority are due to organic
brain damage of prenatal or perinatal origin or to acute metabolic disturbances. The role of neonatal hypoxic-ischemic encephalopathy is probably less than formerly thought because of improvements in prenatal and obstetric care. Moreover, a significant proportion of seizures attributed to anoxia are of the subtle type and may not be epileptic. A relatively poor correlation has been found between such variables as Apgar scores, cord pH, and abnormalities of cardiac rhythms and the occurrence of convulsions.²⁰ Conversely, emphasis has been put more recently on the etiologic role of brain damage of prenatal origin²⁷ and congenital brain anomalies.