ABSTRACT
This paper presents a new measure of heart rate variability (HRV) that can be estimated using Doppler ultrasound techniques and is robust to variations in the angle of incidence of the ultrasound beam and the measurement noise. This measure employs the multiple signal characterization (MUSIC) algorithm which is a high-resolution method for estimating the frequencies of sinusoidal signals embedded in white noise from short-duration measurements. We show that the product of the square root of the estimated signal-to-noise ratio (SNR) and the mean-square error of the frequency estimates is independent of the noise level in the signal. Since varying angles of incidence effectively changes the input SNR, this measure of HRV is robust to the input noise as well as the angle of incidence. This paper includes the results of analyzing synthetic and real Doppler ultrasound data that demonstrates the usefulness of the new measure in HRV analysis
Introduction
Eight out of one-thousand live-born infants have some form of heart defect, making it the single most common class of congenital abnormalities. Identification of these cases during early pregnancy reduces risks by timely treatment and /or planned delivery at tertiary care centers. Currently, the heart rate variability (HRV) analysis is used to understand the hemodynamic state in adults and children. The use of fetal heart rate as a tool for monitoring fetal hemodynamics was first suggested in the 1950s using electrocardiographic (ECG) signals. Unfortunately, there is no reliable or noninvasive method for measuring fetal ECG during early stages of pregnancy. Fetal ECG signals can be measured externally from electrodes placed on the maternal abdomen. However, these signals are in general corrupted by the maternal ECG and other signals such as abdominal action potentials. They are also significantly attenuated by tissues between the heart and the electrodes. The poor signal-to-noise ratio (SNR), and the high rate of coincidence of maternal and fetal ECG’s limits the detection rate of fetal QRS complexes from these composite signals to about 60%. Furthermore fetal ECG can be measured using this method only after 20 weeks of gestation. Fetal ECG signals can also be measured directly from the fetal scalp. However, this method is not feasible until later in gestational age and is performed after the membrane rupture. This paper is concerned with developing a robust measure HRV from Doppler ultrasound measurements of fetal blood flow velocity waveforms. Ultraasonography is a safe, noninvasive, and cost-effective tool for monitoring fetal cardivascular system through imaging and blood flow velocity measurements. Splunder, et al. have shown that blood flow velocity waveforms can be measured from fetal arteries as early as the eighth week of gestation using transvaginal Doppler ultrasound techniques. Short-term temporal and spectral, variability of fetal heart rate have been used for the assessment of cardiovascular development in fetuses during early human pregnancy using the umbilical arterial Doppler ultrasound blood flow measurements. Instantaneous fetal heartbeats were estimated from these blood velocity waveforms using a threshold detection scheme and spectral dynamics of beat-to-beat variability were characterized using fast Fourier transform (FFT) techniques. It is well known that the variance of the spectrum estimate based on the FFT of a random signal (periodogram) is of the order of the spectrum of the signal. Furthermore, the spectral resolution of FFT-based techniques is inversely proportional to the length of the data segments. Both these factors contribute to the unreliability of FFT-based algorithms in HRV analysis. Exposure to ultrasound beam for long durations of time may cause temperature increase and damage in fetal tissues especially in the brain encased in the fetal skull. Consequently, long-term monitoring of blood velocity waveforms is not recommended. Other soruces of error in HRV estimation from blood flow velocity waveforms measured using Doppler ultrasound techniques are: 1) variations in the angle between the incident ultrasound beam and the blood flow: 2) the nonuniform insonation of the vessel: 3) the high-pass filtering of the received ultrasound signal (to demodulate the Doppler signal); and 4) the SNR of the Doppler signal. These factors make it difficult todefine a specific reference point in each pulse in the blood velocity waveform to estimate individual heartbeats and, hence, the beat-to-beat variability.This paper presents a solution to the problems in characterizing the HRV because of variations in the angle of insonation and noise levels in the Doppler signal described above. We used maximum blood flow velocity waveforms estimated from Doppler measurements in our analysis. These waveforms are not significantly affected by the nonuniform insonation and the high-pass filter settings. We first demonstrate that blood velocity waveforms can be modeled as a single sinusoid embedded in white noise over short intervals of time. This model allows us to estimate the heart rate as the instantaneous frequency of the sinusoid in the short data interval. We then define the HRV as the temporal variability of the estimated heart rates over all the short intervals into which the blood velocity waveform is segmented. We use the MUSIC algorithm to estimate the frequency and amplitude of the signal as well as the variance of the noise component in the signal. MUSIC is an eigenanalysis-based method, the spectral resolution of the estimates is not limited by the duration of the signal in this approach. Finally, we use a normalization of the temporal variations in the estimated parameters to develop a measure of HRV that is robust to the ambient noise and the angle of incidence of the ultrasound beam.The rest of the paper is organized as follows. We describe the robust measure of HRV is Section II. Experimental results verifying the robustness properties of the measure is provided in Section III. The Concluding remarks are given in section IV.

download full report

http://ncbi.nlm.nih.gov/pubmed/12892322
http://ieeexplore.ieeexpls/abs_all.jsp?arnumber=1213847