Department of Electrical Engineering, Universidad Autónoma Metropolitana-Iztapalapa, Av. San Rafael Atlixco 186, Col. Vicentina, Alcaldía Iztapalapa, C.P. 09340, México City, México.
Physiol Meas. 2021 Mar 11;42(2):025004. doi: 10.1088/1361-6579/abe0fe.
This work explored the reliability of using points on the heart sounds envelope as indicators of the opening and closure of the aortic valve (AVO, AVC) to measure the pre-ejection period (PEP) and the left ventricular ejection time (LVET).
36 phonocardiograms (PCGs) from healthy subjects and cardiovascular disease subjects were denoised using single-channel independent component analysis (SCICA) and, from the Hilbert envelopes, the positions of the S1 and S2 peaks were detected (pS1, pS2). Complementarily, the positions of the local maxima of S1 and S2 (mS1, mS2) and the points surrounding pS1 and pS2 (tS1, tS2) were obtained. Finally, the reliability of these points (and the corresponding PEP and LVET intervals) was evaluated by the calculation of three error indexes (ePEP, eLVET, and score) and by comparison to reference annotations provided by echocardiography using the Bland-Altman analysis and the paired T-test.
The results indicated that, from a total of 920 and 341 heartbeats in the healthy and diseased groups, respectively, the timing points given by pS1 and pS2 (or mS1 and mS2) were unlikely to substitute for the reference annotations and, thus, are unreliable for measuring the PEP and LVET intervals in the PCG. The t-points evaluation, on the other hand, indicated that tS1 was likely to substitute for AVO and was thus reliable for measuring the PEP using the PCG, with median and interquartile ranges of 0.3(8.3) ms and -0.2(7.5) ms for each group. Future work will generate an envelope with higher temporal resolution, from where tS1 and tS2 can be more accurately detected to improve the PEP and LVET measurements on a larger dataset.
The statistical tests revealed that the envelope of S1 is suitable for extracting a timing point from which the pre-ejection interval can be reliably quantified, and discarded the local maximum used in other studies.
本研究旨在探讨利用心音包络上的点作为主动脉瓣（AVO、AVC）开闭的指标来测量射前期（PEP）和左心室射血时间（LVET）的可靠性。
使用单通道独立分量分析（SCICA）对 36 例健康受试者和心血管疾病受试者的心冲击图（PCG）进行去噪，并从 Hilbert 包络中检测到 S1 和 S2 峰值的位置（pS1、pS2）。此外，还获得了 S1 和 S2 的局部最大值（mS1、mS2）以及 pS1 和 pS2 周围的点（tS1、tS2）的位置。最后，通过计算三个误差指标（ePEP、eLVET 和得分）以及通过与超声心动图提供的参考注释进行 Bland-Altman 分析和配对 T 检验，评估这些点（以及相应的 PEP 和 LVET 间隔）的可靠性。
结果表明，在健康组和疾病组中，分别有 920 次和 341 次心跳，pS1 和 pS2（或 mS1 和 mS2）给出的定时点不太可能替代参考注释，因此，在 PCG 中测量 PEP 和 LVET 间隔是不可靠的。另一方面，t 点评估表明，tS1 可能替代 AVO，因此可用于通过 PCG 测量 PEP，每组的中位数和四分位距分别为 0.3(8.3)ms 和-0.2(7.5)ms。未来的工作将生成一个具有更高时间分辨率的包络，从中可以更准确地检测到 tS1 和 tS2，从而在更大的数据集上提高 PEP 和 LVET 的测量精度。
统计检验表明，S1 的包络适用于提取一个时间点，从该时间点可以可靠地量化射前期，并且摒弃了其他研究中使用的局部最大值。
