Measurement of Duration, Energy of Instantaneous Frequencies, and Splits of Subcomponents of the Second Heart Sound

This paper presents an approach to measure the duration and the energy of instantaneous frequencies (EIFs) of the aortic (A2) and pulmonic (P2) valve closure sounds for the second heart sound (S2) based on analytic signal representation. In past studies, concepts were usually surrounded around the measurement of splits (i.e., delays between the A2 and the P2) in medical terms based only on visual inspections of the time-frequency representation of the S2s. The values were empirically estimated, and the two vital issues—A2 and P2 overlaps and low energies of P2s—were ignored. In this paper, such issues are addressed, and the relevant parameters are measured by identifying the starting and the ending positions of A2s and P2s. The diagnosis related to the duration and the EIFs of the A2 and the P2 is also examined. Furthermore, the proposed method explicitly guides to distinguish the normal/abnormal S2s and the types of S2 splits. The proposed method is characterized by the Hilbert transform-based IF estimation as well as the localization technique based on the reassignment of smoothed pseudo-Wigner–Ville distribution. To validate the idea, total 31 S2s collected from six healthy normal subjects were tested. The result computed by the proposed method shows that the mean ± standard deviation (SD) of the duration of A2s and P2s is 46.7 ± 2.5 and 41.8 ± 2.4 ms, respectively. The mean ± SD of the EIFs of A2s and P2s is 13.8 ± 2.4 and 10.5 ± 1.7, respectively. These estimated results accord with the evidence in the literature. Moreover, when the proposed method was applied to the abnormal S2s, collected from the database of the Texas Heart Institute, it could correctly distinguish the types of abnormal splits. The experimental result reveals that the accuracy rate of the proposed method is as high as up to 97%, thereby demonstrating the effectiveness of the proposed idea.