It is now possible to make a good-quality, continuous electroencephalogram (EEG) record of sick newborn babies receiving intensive care, from which seizure activity can be identified readily (Eyre et al. 1983a, b). The background EEG can be difficult to interpret, however, because it changes with increasing gestational age. In addition, there are cyclical changes in the EEG during sleep that are dependent not only on the baby's sleep state but also on the length of the preceding sleep epoch (Roffwarg et al. 1966, Eyre 1984. Therefore the normal pattern of the EEG cannot be described fully from intermittent, short records. The aim of this study was to describe the range of normal EEG activity at different gestational ages during the first five days after birth by analysing continuous recordings.
Subjects and methodThis study of normal babies was approved by the Central Oxford Hospitals' Ethical Committee and informed consent was obtained from the parents. 25 babies were studied, 12 girls and 13 boys. They were all healthy, with no medical problems during the first five days. Their gestational ages ranged from 26 to 42 weeks and their birthweights from 990 to 4120g. All had a normal neurological examination and cerebral ultrasonogram at the beginning and end of the recording. All were examined at a postconceptional age of 18 months by an independent investigator and were found to be within the normal range for development, with no abnormal neurological signs.A small battery-powered four-channel tape-recorder was used to record two channels of EEG, one of electrocardiogram (ECG) and one of transthoracic impedence. The EEG record was bi-polar, F3-P3, and F4-P4, with an indifferent electrode placed on the scalp. The filter band-width was 0.5 to 1oc)Hz. A more detailed description of the recording technique has been published (Eyre et al. 1983a). The record was started within the first 24 hours after birth and continued, where possible, for five days. The EEG was analysed using two methods: ( 1 ) replay on an automatic visual-display unit with an amplification of 50pv/cm at 15cm/sec; and (2) the data were transcribed onto paper, using an inkjet recorder at a speed of lmm/sec and an amplification of 10pv/mm. EEG records associated with movement on the ECG or transthoracic impedance record were excluded from further analysis. The remaining EEG was subdivided into two distinct patterns, trace-alternant activity and continuous activity. Trace-alternant m 03 z 599