Stringlike Pulse Quantification Study by Pulse Wave in 3D Pulse Mapping

Abstract

Background:

A stringlike pulse is highly related to hypertension, and many classification approaches have been proposed in which the differentiation pulse wave (dPW) can effectively classify the stringlike pulse indicating hypertension. Unfortunately, the dPW method cannot distinguish the spring stringlike pulse from the stringlike pulse so labeled by physicians in clinics.

Design:

By using a Bi-Sensing Pulse Diagnosis Instrument (BSPDI), this study proposed a novel Plain Pulse Wave (PPW) to classify a stringlike pulse based on an array of pulse signals, mimicking a Traditional Chinese Medicine physician's finger-reading skill.

Results:

In comparison to PPWs at different pulse taking positions, phase delay Δθand correlation coefficient r can be elucidated as the quantification parameters of stringlike pulse. As a result, the recognition rates of a hypertensive stringlike pulse, spring stringlike pulse, and non–stringlike pulse are 100%, 100%, 77% for PPW and 70%, 0%, 59% for dPW, respectively.

Conclusions:

Integrating dPW and PPW can unify the classification of stringlike pulse including hypertensive stringlike pulse and spring stringlike pulse. Hence, the proposed novel method, PPW, enhances quantification of stringlike pulse.

Introduction

A sphygmomanometer provides only two measurements for the entire arterial circulation system: the peak of systolic and diastolic pressure. However, more useful information has been observed in the arterial circulation system recently, such as arterial compliance, peripheral resistance, and wave reflection.^1

–6 Spatial and temporal properties are also needed for analyzing the arterial circulation.^7,8 For this purpose, many noninvasive instruments and analysis methods have been proposed,^9

–12 most of which, however, have focused on the cardiovascular system. On the other hand, recent progress in hemodynamics has attracted lots of attention and has led to more discoveries about the physiology of the arterial circulation system, such as the relations among age, gender, drugs, and the wrist arterial pulse waveforms.

The clinical significance of arterial pressure waveforms has been remarked upon since ancient times. Traditional Chinese Medicine Pulse Diagnosis (TCMPD) has especially stood out among these techniques.^3,13 Traditional Chinese Medicine (TCM) physicians use three fingertips to diagnose a patient's health status. Most TCMPD touches focus on the arterial circulation system.^14
–16 Little effort has been spent by researchers on the exploration of the mystery of TCMPD based on TCM theories, and modern researchers are even less clinically experienced in TCM.^17

–20 The main issue is how to interpret the sensations felt by the physicians' fingertips in terms of wrist arterial pulse waveforms. This study is intended to propose a novel method to transform the wrist arterial waveforms into a readable quantification of the fingertip sensation of physicians by mimicking their finger reading skill.^21

–25

The designed pulse-taking platform comprises one displacement detection system, which mimics or duplicates the pulse-taking displacement of physicians,^22

–25 and one pressure-detection system, which obtains information from the wrist arterial pulse waveforms to the greatest possible extent by duplicating a physician's pulse-taking skill.²¹ The consistent pulse conditions acquired from a physician's fingertip sensations will thus be feasible to display. This research, based on the authors' prior experimental results, intends to propose quantifiable methods by which to quantify traditional TCM pulse conditions, such as the stringlike pulse.

Recent research has revealed that the stringlike pulse is highly related to hypertension,²⁶ and its pulse-taking method is based on pressing with one finger. According to these results, the differentiation pulse wave (dPW) has significant characteristics by which to classify the hypertensive stringlike pulse. However, following the basic theory of TCMPD, the stringlike pulse also exists in the spring season for healthy patients, and is called the spring stringlike pulse. Unfortunately, the dPW method cannot distinguish the spring stringlike pulse from the stringlike pulse so labeled by physicians in clinics. The goal of this proposal is to tentatively propose a novel method, which, based on a three-dimensional pulse map (3DPM), can improve the quantification of the stringlike pulse. The pulse taking procedure for the proposed method is based on simultaneous palpation (SP) with an integrated pulse-taking platform.^{21
–23,27,28}

The stringlike pulse and non–stringlike pulse are classified by plain pulse wave (PPW) and dPW. For the PPW, the matched rate in comparison to the physicians' fingertip sensation is 100% in a hypertensive stringlike pulse, 100% in a spring stringlike pulse, and 77% in a non–stringlike pulse.

Methods

Subjects

There are two kinds of stringlike pulse: the spring stringlike pulse and hypertensive stringlike pulse. According to a basic theory of TCM, the spring stringlike pulse is not an illness index but just reflects the body adjustment for weather changes in the spring season. To obtain the spring stringlike pulse, the acquisition time was set in the spring season (i.e., from about March to May). Participants were adopted as hypertensive stringlike pulse candidates, once the participants were diagnosed by physicians as having hypertension. Spring stringlike pulses (healthy participants) were selected from the R.O.C. Air Force Academy; while hypertensive stringlike pulses (hypertensive participants) were selected from National Cheng Kung University Hospital. Hence, the participant inclusion criteria were the following: 1. Health and pulse condition belongs to the stringlike pulse; 2. Hypertension and pulse condition belongs to the stringlike pulse; 3. Health or hypertension but pulse condition belongs to non–stringlike pulse. The procedures for the experiment were approved by the R.O.C. Air Force Academy and National Cheng Kung University. The approved number was 0970006465. Volunteer participants were recruited through questionnaires from the R.O.C. Air Force Academy and National Cheng Kung University. Finally, the participants included 8 healthy students screening from 50 students in the R.O.C. Air Force Academy, and five hypertension patients. Mean age of participants was 32.5±18.8 (standard deviation) years (range 18–64 years).

Bi-sensing pulse-taking instrument

Pulse-taking depth and finger-reading sensations are mimicked by a proposed pulse-taking platform, the Bi-Sensing Pulse Diagnosis Instrument (BSPDI) shown in Figure 1.^21

–25 It comprises three parts: a displacement detection system, a pulse detection system, and a motorized automatic driving system.

FIG. 1.

Bi-Sensing Pulse Diagnosis Instrument, which can reconstruct the pulse-taking method of physicians, such as pulse-taking depth and finger-reading sensation.

In this report, the pulse-taking procedure adopts SP to obtain wrist artery signals. The pressure of physicians contributes to variations in the strain gauge of the displacement detection system in order to mimic their pulse-taking skill. Three (3) displacements, Fu (superficial), Zhong (middle), and Chen (deep), represent different pulse-taking depths based on the TPNI definition (Three Positions Nine Indicators defined in TCM). The robotic fingers are moved by the motorized automatic driving system on the basis of the recorded displacements to duplicate a TCM physician's pulse-taking skill. The pulse detection system then detects the pulse waveforms as the robotic fingers arrive at the recorded TPNI displacements. Twelve (12) sensing elements in one pulse-taking position, such as Cun, Guan, or Chi, are embodied in the pulse-detection system. Cun, Guan, and Chi are three sections over the radial artery for pulse diagnosis by a TCM physician: the Guan is just central to the radial styloid at the wrist where the tip of the physician's middle finger is placed, the Cun is next to it on the distal side where the tip of the physician's index finger is located, and the Chi is on the proximal side where the tip of the physician's ring finger is placed. The physicians' pulse-taking sensations can then be quantitatively represented as an array of pulse waveforms in a 3DPM format. Therefore, the surface of the pulse conditions can be measured and constructed in a mathematical way.

Experimental protocol

TCM has accumulated abundant clinical experience over a period of thousands of years. The principles involved have been qualitative in nature and have not been well quantified. This research is intended to reconstruct the diagnostic environment of TCMPD, which includes the reconstruction of the pulse-taking method and the representation of physicians' observations in a quantitative way.

All of the participants were asked to refrain from exercising or eating and to rest for at least 2 hours prior to their pulse being taken. Figure 2 shows the experimental protocol with five steps. In step 1, the participants were asked to sit on an adjustable chair. Physicians checked all participants to confirm pulse conditions such as a stringlike pulse or a non–stringlike pulse. In step 2, the physician marked the Cun, Guan, and Chi positions of a participant to show where to place displacement sensors of the displacement detection system. Then the participant's wrist was put under displacement sensors, and a physician pushed his/her three fingertips on top of the displacement sensors to touch the Cun, Guan, and Chi positions of the participant. The displacement detection system obtained the participant's pulse-taking depths at Fu, Zhong, and Chen, while the physician checked and recorded the participant's individual pulse conditions. In step 3, the participant's wrist pulses at the obtained pulse-taking depths (Fu, Zhong, and Chen) were measured by the pulse-detection system of BSPDI. The participant changed the other hand to sample wrist pulse and repeat step 2 to step 3. In step 4, a wavelet algorithm, Daubechies wavelets, was used to remove the baseline wander. A moving average filter was adopted to smooth the sampled signals. In step 5, a 3DPM was formed from 12 pulse signals measured by 12 sensors in an array. Nine 3DPMs were obtained at nine TPNI positions. 3DPM can be formulated as below: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}3DPM [ j , k , t ] = V [ j , k , m , n , t ] _{m = 1 \sim 3 , n = 1 \sim 4} , \tag{1}\end{align*} \end{document}

FIG. 2.

Experiment protocol for the quantification of pulse diagnosis. 3DPM, three-dimensional pulse map.

where 3DPM represents the physicians' finger-reading sensations in terms of an array of sensors; j, the pulse-taking position, such as Cun (j=1), Guan (j=2), or Chi (j=3); k, the pulse-taking depth, such as Fu (k=1), Zhong (k=2), or Chen (k=3); m=1, 2, 3, index of sensing elements perpendicular to the wrist artery; n=1, 2, 3, 4, the index of sensing elements parallel to the wrist arteries; t, the sampling time; and V, the pressure measurements in terms of potentials.

To get V, each acquisition took about 20 seconds, and the sample rate was 100 Hz. An ActiveX 3D graph control is used to obtain the intensity graph of pulse conditions. Twelve (12)-channel signals are fed into the ActiveX 3D graph control in LabVIEW, and then the intensity graph is displayed in Figure 3. Figure 3 shows 3DPM taken at Guan position (j=2) and Fu displacement (k=1). Figure 3A demonstrates 12 pulse signals in the array of sensors, while Figure 3B presents 3DPM at a sampling instant in Figure 3A. The highest amplitude is located at the lower and left corner of 3DPM, which represents a strong beat feeling at the bottom and right corner of the fingertip.

FIG. 3.

An example of a 3DPM (3D pulse map) at j=2 (Guan); k=1 (Fu); m=1, 2, 3, index of sensing elements perpendicular to wrist artery; n=1, 2, 3, 4, index of sensing elements parallel to wrist artery. A. The wrist artery waveform at one array of the 3X4 sensor. V_jkmn represents the pressure measurements of each channel. B. 3DPM at a sampling instant of A.

Definition of hypertensive stringlike pulse by dPW

Although TCM physicians have accumulated a great deal of experience with pulse diagnosis over a period of thousands of years, their diagnostic sensations could not necessarily have resulted in the same conclusions due to different levels of pulse-taking skill and subjective personal impressions. Scientists have put great effort into the quantification of pulse conditions instead of into physician's finger-reading sensations to avoid ambiguity and misunderstanding. The stringlike pulse is now one of the pulse conditions that has been well studied in the literature due to the fact that it is easy to determine and because physicians are in agreement as to outcome. According to the World Health Organization, if a pulse sensation feels straight, long and taut, like a musical string to the touch, it is called a stringlike pulse.²⁷

Research results have revealed that the wrist artery pulse waveform of patients with hypertension has significant characteristics that are related to the stringlike pulse²⁶ in which V_a1 and V_a2 are key parameters for classification of hypertensive stringlike pulse. The tidal wave increases earlier than other pulse waveforms due to the stiffness of the wrist artery and the presence of a high valley in valley “V_a2 ” of the differentiation pulse wave (called dPW), as shown in Figure 4. In this proposal, such a wrist artery pulse waveform is called a hypertensive stringlike pulse.

FIG. 4.

A. The hypertensive stringlike pulse waveform. B. The differentiation pulse wave of A: V_a1 and V_a2 are key parameters for classification of hypertensive stringlike pulse. dPW, differentiation pulse wave.

Definition of the stringlike pulse by PPW

PPW elucidated from 3DPM is defined below: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}\frac{PPW [ j , k , t ] = \mathop \sum \limits_{m = 1}^3 \ \mathop \sum \limits_{n = 1}^4 V [ j , k , m , n , t ]}{12} , \tag{2}\end{align*} \end{document}

where PPW is simply equal to the average of 12 pulse signals in a 3DPM.

PPW is proposed based on the clinic experiences of stringlike pulse by TCM physicians: synchronous pulse beating at least two fingertips. PPW could represent the average beating feeling of the fingertip of a TCM physician. In Figure 5, two PPWs at Cun and Guan positions are shown, in which while “T” is the beating period, “r” is the correlation coefficient, and Δt is the time difference. The r parameter is defined as below: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} r = \frac {\sum ( X - \overline {X} ) ( Y - \overline { Y } )} { \sqrt{ \sum ( X - \overline { X } ) ^2 \sum ( Y - \overline { Y })^2 }} , \tag { 3 } \end{align*} \end{document}

FIG. 5.

r, Δt, and T at two plain pulse wave (PPWs).

where either X or Y is PPW at Cun, Guan, or Chi. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline{X}$$ \end{document} or \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$\overline {Y}$$ \end{document} is the average of PPW.

Δt can be obtained by measuring the time difference of two PPWs' peak, i.e., \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \Delta t [ p , q , k ] = PPW [ p , k , t ] - PPW [ q , k , t ] , \tag{4}\end{align*} \end{document}

where Δt[p, q, k] represents the time difference between two PPWs at p (j=p) and q (j=q) pulse-taking positions at the same k pulse-taking depth.

Time difference Δt can be transformed into phase angleΔθ \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \Delta \theta = \frac { \Delta t [ p , q , k ] } { T } \times 360 ^ \circ , \tag { 5 } \end{align*} \end{document}

Statistical method

The aim of this research is to propose a novel approach to enhance the classified method of stringlike pulse. To acquire a statistical analysis, the SPSS 17.0 program was used. An independent sample t-test and analysis of variance were carried out to check the feasibility of quantification of the stringlike pulse.

Results

Two (2) kinds of stringlike pulses are listed in Huangdi's Internal Classic: one is the pulse characteristic of illness; the other is seasonal normal pulse of spring. Figures 6A, 7A, and 8A demonstrate the general hypertensive stringlike pulse, spring stringlike pulse, and non–stringlike pulse separately. The corresponding dPW and PPW are shown in B and C of Figures 6, 7, and 8 separately. Clearly, dPW cannot classify normal pulse beating in a stringlike format in spring in Figure 7B, where |V_a2 | is quite small. Phase angle Δθ is quite large at non–stringlike pulse in Figure 8C, while Δθ is relatively small at stringlike pulse in Figures 6C and 7C.

FIG. 6.

An example of a hypertensive stringlike pulse. A. Original pulse signal, (B) differentiation pulse wave (dPW) with V_a1 =0.0091 mV/sec and V_a2 =−3.8922 mV/sec, and (C) plain pulse wave (PPW) with r=0.7902 and Δt=60 ms or Δθ=26.34°.

FIG. 7.

An example of a spring stringlike pulse. A. Original pulse signal, (B) differentiation pulse wave (dPW) with V_a1 =−4.0500 mV/sec and V_a2 =−2.5832 mV/sec, and (C) plain pulse wave (PPW) with r=0.4513 and Δt=80 ms or Δθ=27.96°.

FIG. 8.

An example of a non–stringlike pulse. A. Original pulse signal, (B) differentiation pulse wave (dPW) with V_a1 =−5.0246 mV/sec, and V_a2 =−1.9272 mV/sec, (C) plain pulse wave (PPW) with r=−0.7732 and Δt=350 ms or Δθ=168°.

Table 1 summarizes the statistical measurement results by dPW (V_a1 and V_a2 ) and PPW (r and Δθ) for stringlike pulse (hypertensive and spring) and non–stringlike pulse. In spring stringlike pulse, |V_a2 | is insignificantly larger than |V_a1 |, so dPW fails to classify spring stringlike pulse. However, dPW significantly recognizes hypertensive stringlike pulse in comparison to spring stringlike pulse and non–stringlike pulse. r and Δθ of PPW significantly classifies between stringlike pulse (including hypertension and spring) and non–stringlike pulse.

Table 1.

Summary of Quantification Parameters of Differentiation Pulse Wave and Plain Pulse Wave for Hypertensive and Spring Stringlike Pulse and Non–Stringlike Pulse

	Stringlike pulse
	Hypertensive	Spring	Non–stringlike pulse	p-Value
V_a1 (mV/sec)	−1.2329±0.7688	−1.8036±0.8854	−1.6935±1.0850	0.196
V_a2 (mV/sec)	−2.1507±1.9662	−1.0596±0.5489	−1.2342±0.9242	0.013
r	0.4856±0.2726	0.6162±0.2638	−0.3512±0.1318	0.000
Δθ (Degree)	35.0266±20.6572	24.2896±16.480	141.2334±41.3008	0.000

Discussion

This article is intended to provide a novel approach by which to explore subjective pulse conditions based on traditional TPNI pulse-taking methodology. The stringlike pulse is easily found in patients with hypertension and during the spring season. The hypertensive stringlike pulse shows its tidal wave approaching the percussion wave due to the fact that systolic pressure augments mostly in elderly patients.³ In contrast, “The spring stringlike pulse normally appears in healthy persons in spring,” is described in Huangdi's Internal Classic. Apparently, these two kinds of stringlike pulses have very different waveforms, as shown in Figures 6A and 7A. In Figure 9, p-value is applied among hypertensive stringlike pulse, spring stringlike pulse, and non–stringlike pulse for significance judgment. The V_a2 of dPW is significant with regard to distinguishing a hypertensive stringlike pulse from the spring stringlike pulse and the non–stringlike pulse, but insignificant between spring stringlike pulse and non–stringlike pulse, as depicted in Table 1 and Figure 9A. To classify stringlike pulse by PPW with synchronous beating characteristic, the higher r and lower Δθ yield better synchronicity of the stringlike pulse. In Figures 9(b) and 9(c), r and Δθ are significant to classify stringlike pulse including hypertensive and spring from non–stringlike pulse, but insignificant to recognize between hypertensive and spring stringlike pulse. Apparently, PPW is able to recognize stringlike pulse but PPW needs dPW to help to distinguish between hypertensive and spring stringlike pulse. In summary, r and Δθ of PPW are used to distinguish a stringlike pulse (hypertensive or spring stringlike pulse) from a non–stringlike pulse, while V_a2 of dPW is adopted to differentiate between a hypertensive stringlike pulse and a spring stringlike pulse.

FIG. 9.

Bar-plot for significance analysis among hypertensive, spring, and non–stringlike pulses. (S, significance as p<0.05; N.S., no significance as p>0.05). A. V_a2 for differentiation pulse wave. (B) r for PPW. (C) Δθ for plain pulse wave.

To calculate the recognition rate of dPW and PPW in comparison to the physicians' fingertip sensations, a stringlike pulse is calculated by setting r>0.1 and Δθ<90° for PPW and |V_a2 |>|V_a1 | for dPW. Table 2 summarizes the recognition rates of PPW and dPW. The PPW recognition rate is 100% for the hypertensive stringlike pulse, 100% for the spring stringlike pulse, and 77% for the non–stringlike pulse. dPW can classify the hypertensive stringlike pulse at a 70% recognition rate but be unable to recognize the spring stringlike pulse. The recognition rate of the non–stringlike pulse is 59% for dPW, lower than the 77% for the PPW.

Table 2.

Summary of Recognition Rates

	Stringlike pulse
	Hypertensive (N=10)	Spring (N=12)	Non–stringlike pulse (N=13)
dPW	70%	0%	59%
PPW	100%	100%	77%

dPW, differentiation pulse wave; PPW, plain pulse wave.

Table 3.

Summary of Abbreviations

Abbreviation	Definition
BSPDI	Bi-Sensing Pulse Diagnosis Instrument
dPW	Differentiation Pulse Wave
PPW	Plain Pulse Wave
PWOF	Pressing with One Finger
SP	Simultaneous Palpation
TCM	Traditional Chinese Medicine
TCMPD	Traditional Chinese Medicine Pulse Diagnosis
TPNI	Three Positions Nine Indicators
3DPM	3-Dimensional Pulse Map

As a result, PPW provides a better recognition rate of stringlike pulse than dPW, as shown in Table 2. However, dPW effectively separates hypertensive stringlike pulse from spring stringlike pulse as shown in Table 1 and Figure 9. In conclusion, PPW does effectively enhance the recognition of stringlike pulse from non–stringlike pulse in addition to dPW.

Conclusions

The aim of this proposal is to enhance a quantification approach to the stringlike pulse by using reconstruction of the pulse-taking environment. The reconstruction includes two aspects: one is to mimic the pulse-taking depth of the physicians; the other is to mimic their fingertips sensations. Only with the implementation of these two conditions will the mysteries of TCMPD be accessed. Hence, this research proposes a bi-sensing pulse diagnosis instrument with an array sensor to be applied to discover ways to quantify the stringlike pulse. The V_a2 , r, and Δθ are found to classify the stringlike pulse. The characteristics of the stringlike pulse are significant, with higher r and lower Δθ, not vice versa. The application of a 3DPM to analyze the pulse conditions provides a novel and feasible approach for the quantification of stringlike pulse. The authors believe that by utilizing this method, more will be known about the human body and the abundant assets of TCM will be comprehended and applied.

Footnotes

Acknowledgments

This work is supported by the National Science Council of Taiwan under grant number NSC-99−2221-E006-109. The authors would like to thank PPS Company for acquisition software and the R.O.C. Air Force Academy for providing test subjects.

Disclosure Statement

The authors declare that they have no competing interests.

References

, Wang

, Hu

, Zhang

. The relationship between pulse wave velocity and pulse pressure in Chinese patients with essential hypertension. Hypertens Res, 2003; 26:871–874.

Klimashow

. Intelligent pulse express diagnosis of the state of human cardiovascular system. Biomed Eng, 2008; 42:284–288.

Karamanoglu

. A system for analysis of arterial blood pressure waveforms in humans. Computers Biomed Res, 1997; 30:244–255.

Wei

, Lo

, Wu

. Spectral analysis of blood circulation based on the viewpoint of resonance. Jpn J Appl Phys Part 1 Reg Papers Brief Commun Rev Papers, 2006; 45:2854–2862.

Song

, Wang

. Pulse wave: The bridge connecting traditional Chinese medicine with Western medicine. Zhong Xi Yi Jie He Xue Bao, 2008; 6:891–896.

Lee

, Wei

. Spectrum analysis of human pulse. IEEE Trans Biomed Eng, 1983; 30:348–352.

Wang

YYL

, Sze

, Bau

et al. The ventricular-arterial coupling system can be analyzed by the eigenwave modes of the whole arterial system. Appl Phys Lett, 2008; 92:3.

Wang

YYL

, Chiu

, Jan

et al. Analysis of transverse wave as a propagation mode for the pressure pulse in large arteries. J Appl Phys, 2007; 102:4.

, Wang

, Zhang

. Modern research on Traditional Chinese pulse diagnosis. Eur J Oriental Med, 2004; 4:46–54.

10.

, Chang

, Jiang

. A novel pulse measurement system by using laser triangulation and a CMOS image sensor. Sensors, 2007; 7:3366–3385.

11.

Tyan

, Liu

, Chen

et al. A novel noninvasive measurement technique for analyzing the pressure pulse waveform of the radial artery. IEEE Trans Biom Eng, 2008; 55:288–297.

12.

Chen

, Zhang

. Wrist pulse signal diagnosis using modified Gaussian models and Fuzzy C-Means classification. Med Eng Phys, 2009; 31:1283–1289.

13.

Burns

, Burns

. Western frontier physician taking pulse. J Altern Complement Med, 1999; 5:401–403.

14.

Hua

. Classic of Difficult Issues. Taipei: Publishing House of Wen Guang, 2002.

15.

Wang

. Diagnostics of Traditional Chinese Medicine. Shanghai: Publishing House of Shanghai University of Traditional Chinese Medicine, 2000.

16.

Wang

. The Pulse Classic: A Translation of the Mai Jing. Boulder: Blue Poppy Press, 2002.

17.

Choi

, Kim

et al. Development of spatial pulse diagnostic apparatus with magnetic sensor array. J Magnetism Magn Materials, 2007; 310:E983–E985.

18.

Kim

, Hwang

, Choi

et al. Improvement of pulse diagnostic apparatus with array sensor of magnetic tunneling junctions. J Appl Phys, 2006; 99:3.

19.

Jin

, Yu

. Research of multi-point pulse wave computer measurement system using PVDF. J Tsinghua Univ Sci Technol, 1999; 39:118–121.

20.

Chen

, Liu

. Research on Characteristics of Pulse Delineation in TCM and Omnidirectional. IEEE International Symposium on IT in Medicine and Education. Xiamen: IEEE, 2008; 536–538.

21.

, Chung

, Yeh

, Luo

. Temporal and spatial properties of arterial pulsation measurement using pressure sensor array. Evid Based Complement Alternat Med, 2012; 2012:745127.

22.

, Chung

, Luo

et al. Pulse Differences and 3D Pulse Mapping in TPNI Displacements. 2011 4th International Conference on Biomedical Engineering and Informatics. Shanghai: IEEE, 2011; 830–834.

23.

Chung

, Hu

, Chu

et al. Exploring the Conventional Pulse Conditions Using Bi-Sensing Pulse Diagnosis Instrument. 2011 4th International Conference on Biomedical Engineering and Informatics. Shanghai: IEEE, 2011; 746–750.

24.

Luo

, Chung

, Hu

et al. Possibility of quantifying TCM finger-reading sensations: I. Bi-Sensing Pulse Diagnosis Instrument. Eur J Integr Med, 2012, 10.1016/j.eujim.2012.03.003.

25.

Chung

, Hu

, Luo

et al. Possibility of quantifying TCM finger-reading sensations: II. An example of health standardization. Eur J Integr Med, 2012, 10.1016/j.eujim.2012.03.004.

26.

Fei

. Contemporary Sphygmology in Traditional Chinese Medicine. Beijing: People's Medical Publishing House, 2002.

27.

WHO. WHO International Standard Terminologies on Traditional Medicine in the Western Pacific Region. New York: Renouf Publishing Co. Ltd, 2007.

28.

Zheng

, Li

. Complete Collection of Zhou Xue Hai Medical Works. Beijing, China: China Press of Traditional Medicine, 1999.