Blood Pressure Estimation using Sech Functions Pulse Decomposition

摘要

Blood pressure (BP) is one of the most important indicators of physical and mental health. BP monitoring helps with controlling ailments such as heart disease and may help with stress assessment. Currently, BP measuring technologies use inflatable cuffs, which are inconvenient to use, being undesirable for continuous BP monitoring. Here we use photoplethysmographic (PPG) pulse wave contour to estimate BP using pulse decomposition analysis. As blood is ejected from the heart, the pulse moves both distally to the arms and down the aorta, partially reflecting at the renal and iliac arteries branches. These reflections move up the aorta and distally to the extremities, such that finger PPG signals are composed of three waves: the primary pulse and two delayed reflected waves. This model has been used by others, fitting Gaussian waves to the PPG signal. However, fitting stability and correlation with BP could be improved. In our proposed method, each PPG pulse is the sum of three hyperbolic secants (sech) waves, whose features are determined by PPG curve fitting. An increase in blood pressure makes pulse wave velocity increase, decreasing the intervals between waves. To verify the method, we have collected PPG signals and continuously measured BP from a volunteer. Multiple regression analysis between the PPG extracted features and continuous BP readings shows a very high correlation and high statistical significance. The decomposition in sech showed both higher stability and better correlation with BP than the Gaussian wave decompositions reported in the literature.