A Low Frequency Oscillation in Pulmonary Circulation and Its Dynamic Relation to the Low Frequency Oscillation of Systemic Circulation : Power Spectrum and Phase Estimation by Autoregressive Algorithm and Cross Spectral Analysis

Yum, Myung Kul; Kim, Nam Su

Abstract

BACKGROUND
Low frequency oscillation of systemic artery pressure was known as the marker of sympathetic modulation. Recently the low frequency oscillation of pulmonay artery pressure in pulmonary hypertensive patient was reported. But no further study about its quantitative relationshop and phasic coupling between the low frequency oscillation of pulmonary artery pressure and systemic artery pressure. Power spectral analysis with autoregressive algorithm and cross spectral analysis are powerful tool for investigation these relationship. METHOD: Analog signals of simultaneous measured left pulmonary and femoral artery pressure in thirty one patients with ventricular septal defect were digitized and stored. After modeling each time series with autoregressive algorithm, power spectral density function was obtained by calculation the frequency response function of each model, and then low frequency power was computed. Cross spectral density function provided squared coherence and phase spectrum. Phase between the low frquency oscillation of the two signal was measured from the phase spectrum when the squared coherence is above 0.5.
RESULTS
The advantage of using autoregressive model was that the power spectral density function was continous and sharp spectral peak was usually found. In patients with Rp/Rs<0.25, the low frquency power of pulmonary artery pressure(12+/-12) was significantly smaller than that of the systemic artery pressure(144+/-242). In patients with Rp/Rs>or=0.25, there was no significant difference between the low frequency power of pulmonary artery pressure(384+/-461) and that of the systemic artery pressure(752+/-1241). In patients with Rp/Rs>or=0.25, it was more probable that low frequency oscillation of pulmonary and systemic artery pressure was timely coherent(sqaured coherence>0.5) than in patient with Rp/Rs<0.25. And their phase difference was 0~1.96 radian.
CONCLUSION
Autoregressive algorithm is a more powerful tool for spectral analysis than the method of conventional spectrum estimation. When pulmonary vascular resistance remains low, the low frequency oscillation of pulmonary artery pressure was negligible compared to systemic artery pressure. But as pulmonary vascular resistance elevates, the low frequency power of pulmonary artery pressure is much the same as that of systemic artery pressure, and there is a explicit time realtionship that pulmonary artery pressure leads the systemic artery pressure about 0~3 seconds in the low frequency range.