Pulse Signal Analysis of Patients with Coronary Heart Diseases Using Hilbert-Huang Transformation and Time-Domain Method

• Published in: Chinese journal of integrative medicine Vol. 21; no. 5; pp. 355 - 360
• Main Author: 李福凤 孙仁 薛松 钱鹏 燕海霞 张伟妃 孙祝美 张红凯 王忆勤
• Format: Journal Article
• Language: English
• Published: Beijing Chinese Association of Traditional and Western Medicine 01.05.2015
• Subjects: Adult; Algorithms; Coronary Disease - diagnosis; Female; Humans; Male; Medicine; Medicine & Public Health; Middle Aged; Original Article; Signal Processing, Computer-Assisted; Time Factors; 信号分析; 信号采集系统; 冠心病; 希尔伯特; 患者; 时域法; 脉冲信号; 黄变; time domain; coronary heart diseases; pulse signal; Hilbert-Huang transformation
• ISSN: 1672-0415; 1993-0402; 1993-0402
• DOI: 10.1007/s11655-015-2173-7

Objective： To explore characteristic of pulse signal to distinguish patients with coronary heart diseases （CHD） from patients without CHD and healthy adults, and accordingly evaluate the potential role of pulse signal to diagnosis CHD. Methods： Totally 407 patients enrolled from 4 collaborating medical centers were assigned to a CHD group （205 patients） and a non-CHD group （202 patients）. The healthy control group （62 adults） enrolled from Shanghai University of Traditional Chinese Medicine. Pulse signals were collected using the synchronous multiplex pulse signal acquisition system. The pulse signals were analyzed and extracted using Hilbert-Huang transformation （HHT） and time-domain, respectively. The time-domain parameters of pulse signal were processed by the analysis of variance （SNK test）. Results： Special patterns in the CHD group pulse signal were found in this study： （1） time-domain parameters of pulse signal, h1 , h3, h4, h3/h1, t, t5/t4 were increased and w was wider; （2） 44% of C2 waves in HHT were chaotic and disordered and 72% of C3 waves were exhibited irregularly with average amplitude over 10 g-forces, which were all significantly different from controls. Conclusion： Characteristic wave and time-domain parameters of pulse signal were extracted using HHT and time-domain which could be served as a non-invasive approach for assessing patients with CHD.