One of the highest risk periods for seizures during the human life span is the first month of life. Most neonatal seizures are triggered by acute illness such as hypoxic-ischemic encephalopathy, stroke, or infection; rarely are they triggered by epilepsy per se. Seizures are the most common and important sign of acute neonatal encephalopathy, are a major risk for death or subsequent neurologic disability, and by themselves may contribute to an adverse neurodevelopmental outcome. Customary clinical practice includes visual monitoring of high-risk neonates for seizures, performance of a routine electroencephalogram (EEG) for suspicious clinical seizure activity, and empirical treatment with phenobarbital. Presently, however, there are no data that have unequivocally demonstrated the efficacy of barbiturates in the treatment of neonatal seizures. The neurology group recognizes an important need for randomized, placebo-controlled, ethically acceptable trials of phenobarbital efficacy and safety in the treatment of neonatal seizures. After exploring 3 possible frameworks for clinical trials of phenobarbital in the treatment of neonatal seizures, the neurology group ultimately focused on a multicenter, placebo-controlled, electroencephalographer-blinded study of phenobarbital versus placebo in a homogeneous group of newborns who are at high risk of developing early subclinical electroencephalographically detected neonatal seizures. Continuous video-EEG monitoring would establish the presence and number of seizures. Criteria for escape from the study to treatment are clearly defined. The proposed study could provide the first concrete evidence of treatment efficacy because (1) it examines a homogeneous patient population, (2) the recognition and quantification of seizures rests solely on the gold standard of seizure detection (EEG), and (3) the dosing of phenobarbital is matched specifically to the phenobarbital-binding characteristics of the individual treated. This study would affirm or refute the common practice of phenobarbital as the first-line treatment of neonatal seizures.