A PC based neural network algorithm for measurement of heart rate variability

Heart Rate Variability has recently been shown as a viable index to predict sudden cardiac death. The goal of this research is to investigate the use of neural network technique to classify detected QRS complexes into normal and abnormal ones. A single layer perceptron neural network is used for this QRS pattern learning and classification. Results with real data showed that the algorithm gives a 99% correct QRS detection rate.