In recent years, prolonged EEG monitoring with simultaneous observation by video recorder on by direct inspection has produced findings that suggest that (1) the number of certain neonatal seizures (when identified only clinically) has been overestimated in the past, and (2) the total number of neonatal seizures (identified electrographically but clinically silent) has been underestimated in the past. In this article, we propose a classification of seizures that includes the following as major seizure types: subtle, clonic, tonic, and myoclonic. Subtle seizure phenomena include those alterations in neonatal behavior, motor functions, and autonomic function that are easily overlooked and that are not characterized by clonic, tonic, or myoclonic activity. Such seizures include certain ocular phenomena, oral-buccal-lingual movements, peculiar limb movements, autonomic alterations, and apnea. Subtle seizures appear to be more common in premature than in full-term infants, and some subtle clinical phenomena in full-term infants are not associated with simultaneous EEG seizure activity. Clonic seizures include focal and multifocal varieties—both are accompanied by simultaneous EEG seizure activity. Tonic seizures include focal episodes (less common) and generalized episodes (more common). Generalized tonic seizures mimic decerebrate and decorticate posturing and are not consistently accompanied by EEG seizure activity. Focal tonic episodes are consistently accompanied by such electrographic activity. Myoclonic seizures may be focal, multifocal, or generalized. Only the last of these is commonly accompanied by EEG seizure activity. Thus, the clinical seizure types commonly associated with EEG seizure activity are certain subtle seizures, focal and multifocal clonic seizures, focal tonic seizures, and generalized myoclonic seizures.
Neonatal clinical seizures not accompanied by EEG seizure activity may represent movements or posturing generated by diencephalon-brainstem, "released" from the inhibitory influence of cerebral cortex. This hypothesis is supported by the common occurrence of such clinical phenomena in infants with severe bilateral cerebral injury and certain experimental findings in animals subjected to decortication.
The possibility that neonatal clinical phenomena not accompanied by EEG seizure activity are nevertheless epileptic in origin is raised by documentation of apparent epileptic phenomena in older patients in the absence of surface-recordable electrographic seizure, Certain observations in human newborns and in neonatal animals also suggest this possibility. Further study is needed to resolve this issue.
Distinction of epileptic and nonepileptic phenomena can be made at the bedside in many cases. Thus, nonepileptic phenomena generally are provoked or exacerbated by sensory stimulation, are suppressed by passive restraint, and are not accompanied by autonomic phenomena; the opposite is true for epileptic phenomena.
The issues raised have important implications for management—in particular, decisions concerning whom to treat, criteria for determining adequacy of therapy, and determination of duration of therapy. It has become increasingly clear that most infants with neonatal seizures require only relatively brief treatment with anticonvulsant medications.