Stability of the Frequency Spectrum of the Heart Sounds S1 and S2 under Different Physiological Conditions

Abstract

This paper presents a study of the frequency spectrum regarding the first and the second heart sounds. Therefore, measurements of the phonocardiogram under different physiological conditions are conducted. For this purpose, two different methods to analyze the spectrum, are applied to the recorded phonocardiograms, namely the short-time Fourier transform and the continuous wavelet transform. Seven different measurements are conducted from 12 subjects, where the auscultation point, posture, breathing and physical stress are varied. Moreover, the interference due to human voices is taken into account. The determined 10 dB-bandwidth, which starts at around OHz, for the first and the second heart sound is approx. 55 Hz in the short-time Fourier case and about 100Hz for the continuous wavelet transform. The analyzed parameters do not show any significant change under varying physiological conditions, whereas the frequency spectrum of the first and second heart sounds differs strongly between the individual subjects. It is concluded that for the filter design of heart sounds monitoring in wearable devices, adaptation to varying daily activities can be neglected, whereas adaptation to the user is strongly recommended.