Real Time Analysis of Fetal Heart Rate Variability Utilizing Doppler Ultrasound

The phenomenon of fetal heart rate variability is well known in medicine.

Abstract
The phenomenon of fetal heart rate variability is well known in medicine.
Having a fetal beat-to-beat heart rate, it is possible to perform its spectral analysis. The results of this spectral analysis allow prediction of certain infant diseases and estimation of the general health condition of the fetus.
Taking a pure high quality signal from fetus heart is a problem. It is common practice to measure mechanical signals of fetus heart rate using Doppler Ultrasound. Obtaining good quality signals from fetus heart rate is a very difficult task, since they depend on many external factors (breathing or moving of the fetus, noise, mother signals, etc). Therefore a new signal-processing algorithm was needed to overcome this problem.
This project, based on former works, improves the algorithm and runs an online program that samples the Ultrasound signal and presents the heart rate signal in realtime to the operator.
The project utilizes a PC with a sampling card and Labview as its development environment.
Labview is a highly procedural graphic programming environment that combines easy to use graphical programming with the flexibility of a powerful programming language.

Basic approach
The method used in this project for beat-to-beat heart rate estimation is based on auto-correlation and cross-correlation.
The following block diagram presents the basic algorithm steps:

1

Build frame – Builds an array that contains about 8 fetus heart beats.
AC Calc – Calculates the positive part of the Auto-correlation’s frame.
Find Average Heart Rate – Estimates the average heart beat in the frame using two methods AGM & GLM.
Perform AGT – Performs an auto-correlation quality test by finding 2 more peaks in the auto-correlation signal.
Calculate CC – Calculates cross-correlation between 1 period of the signal and the whole frame.
Find beat-to-beat – finds the actual fetus heart beat.

Results
2

Conclusions
The algorithm was found to be efficient and reliable. It has been tested on over 50 different fetus heart rate samples. The results of the algorithm were compared to the output of the Ultra sound machine and a high degree of correlation has been found.
The algorithm succeeds to provide results in real time.
The application is easy to use with simple MMI (Man Machine Interface) and enables to restore and analyze samples offline.

Acknowledgments
We would like to thank everybody who assisted us in carrying out our project:
Prof Dan Adam, Dr. Israel Taller, and special thanks to B.Sc Johanan Erez and the PSPL laboratory stuff.
The Ollendorff Research Center Fund supported this project.