Estimating Cardiac Stroke Volume from the Seismocardiogram Signal

Abstract

Seismocardiography (SCG) provides a record of mechanical vibrations created by the beating heart and the ejection of blood into the aorta. This signal can be recorded from the surface of the sternum with accelerometers and contains information related to the amount of blood expelled from the heart (stroke volume). The goal of this study was to compare the accuracy of estimation of stroke volume from SCG compared to Doppler ultrasound method. The ultrasound measurement is based on the fact that the velocity of the blood ejected from the heart through the aorta causes a Doppler shift in the frequency of the reflected ultrasound waves. †The integration of the velocity over a single beat combined with measurement of aortic diameter provides an established method for estimation of beat-to-beat stroke volume.

Ten healthy male subjects participated in this study and SCG signals were recorded from them using an accelerometer from PCB Piezotronics, model 393C and the simultaneous Doppler ultrasound signal was recorded by directing the 2 MHz probe (Multi-Flow, DWL Gmbh) towards aorta. †The Ultrasound sample was taken from a position within the aorta approximately 10 mm above the aortic valve. †A standard 3-lead electrocardiograph was used to provide and ECG tracing for identification of individual beats.

Features from the SCG signal were extracted and the simultaneous ECG signal was used to identify every single beat. The extracted features were fed to a feed forward neural network trained by back propagation method and the preliminary estimation results showed that the stroke volume values obtained by SCG signal closely follows the values derived from Doppler ultrasound method. Bland and Altman method was used to compare the two methods.