Artificial intelligence based classification of menstrual phases in amenorrheic young females from ECG signals

In the present study, attempts were made to classify menstrual phases of young healthy female (21-25 years) based on the features obtained from ECG signals. Statistical features were extracted from the heart rate variability (HRV) and the ECG signals and were used for pattern recognition during the different menstrual phases. The pattern recognition studies using HRV features suggested that the menstrual phase classification efficiency were >85 % and > 90 % using Multilayer perceptron (MLP) and Radial basis function network (RBF) Artificial Neural Network (ANN) models. On the other hand, the pattern recognition studies using ECG signal features showed classification efficiencies of > 80 % and > 90 % using MLP and RBF ANN models. The results indicated temporary changes in the autonomic nervous system and the cardiac physiology of the volunteers during the menstrual cycle.