Changes in the VEP in preterm neonates with arousal states, as assessed by EEG monitoring.

The effect of sleep state on the visual evoked potentials (VEPs) in neonates was investigated in 7 preterm infants. Polygraphic monitoring including EEG, EOG, ECG, respirogram and submental EMG for the purpose of sleep staging was carried out on all infants simultaneously with VEP testing. Awake-sleep stages were classified into 4 states: awake, transitional or atypical, quiet sleep and active sleep. VEPs were recorded from Oz, referenced to Fz, in response to binocular stimulation with light-emitting diode goggles. Polygraphic and EEG data were analyzed separately. Reproducible VEPs that were appropriate for their ages were seen in all infants in the awake state. The N300 was the most reliable component across sleep states but there was a significant decrease in amplitude with quiet sleep. There were no significant differences among awake, atypical or active sleep states. The other two VEP components followed the same trends as the N300. The P200, present in the older infants, disappeared in both active and quiet sleep states; the P400 was typically variable but reliably present in the awake or atypical states. When a distinction is made between quiet sleep and other arousal states, consistent and significant differences emerge. Our results emphasize the need to test infants in the same arousal states in studies of VEPs in order to make valid comparisons of latency or amplitude changes, particularly with longitudinal or follow-up studies.