Visser G H, Dawes G S, Redman C W
Br J Obstet Gynaecol. 1981 Aug;88(8):792-802. doi: 10.1111/j.1471-0528.1981.tb01305.x.
A numerical method for separation of the frequency components of the fetal heart rate, and for identifying episodes of high or low variation, was applied to 196 64-minute ultrasound records in normal pregnancies during the last half of gestation. Most variables of heart period variation increased with gestation, as did the incidence of accelerations. Before 35 weeks, accelerations of greater than or equal to 14 beats/minute did not occur in all records. Cyclic episodes of low and high variation ("unreactive' and "reactive' episodes), associated with rest-activity cycles, could be identified from 27 weeks onwards. After 36 weeks gestation the length of low variation episodes increased and the variation during these episodes fell. Near term, low variation episodes lasted up to 40 minutes. It is concluded that on scrutiny of fetal heart rate records for evidence of normality, or otherwise, due account should be taken of gestational age.