1. Introduction
Traditional Chinese Medicine (TCM) pulse diagnosis stands as a cornerstone of TCM assessment, integral to consultations for centuries. This sophisticated technique, involving pulse palpation at three distinct locations—cun, guan, and chi—on both wrists, offers a comprehensive understanding of an individual’s overall health and the condition of specific organs. Figure 1 illustrates these key locations and their corresponding organ associations, serving as a vital guide for TCM practitioners. By synthesizing data from pulse assessment with other clinical observations, TCM doctors formulate personalized treatments and meticulously monitor patient progress.
Figure 1.
Distribution of organs at the six locations (Adapted from [1])
The escalating global interest in TCM has propelled Chinese Pulse Diagnosis into the spotlight, sparking widespread curiosity about its scientific underpinnings and clinical efficacy. Since the 1950s, a significant body of research has emerged, dedicated to quantifying chinese pulse diagnosis. This endeavor seeks to establish a robust scientific foundation, thereby validating its clinical significance and bridging traditional wisdom with modern scientific rigor. This review aims to provide an exhaustive overview of the current advancements in the quantification of chinese pulse diagnosis, offering readers a complete understanding of its contemporary progression.
Upon completion of this chapter, readers will be equipped to:
- Grasp the latest scientific evidence supporting the quantification of chinese pulse diagnosis.
- Critically evaluate the strengths and weaknesses of existing research methodologies and statistical approaches in this field.
- Identify future directions for the quantification of chinese pulse diagnosis, paving the way for continued advancement and integration with modern healthcare practices.
This review is structured into five key sections. The initial three sections delve into the qualification and quantification of chinese pulse diagnosis as documented in both ancient and contemporary literature. A dedicated section explores the statistical methodologies employed in quantifying chinese pulse diagnosis. Section four introduces a TCM pulse diagnostic framework, conceived by the author in 2010, to elucidate the intricate relationship between pulse conditions and arterial pulse dynamics. The final section addresses the limitations inherent in current studies and proposes actionable recommendations to guide future research and development in this vital area of TCM.
2. Qualification of TCM Pulse Diagnosis
The qualification of TCM pulse diagnosis pertains to defining the essential elements that constitute a thorough and valid pulse assessment within the TCM framework. Scholarly literature reveals considerable ambiguity and confusion surrounding pulse assessment in TCM, primarily stemming from the often-vague descriptions of pulse conditions found in classical Chinese medical texts [1].
While the pulse itself is an objective physiological phenomenon, the interpretation of pulse condition is inherently subjective. It represents the qualitative attributes of the pulse as perceived by a TCM doctor, reflecting their individual clinical judgment. Over thirty distinct pulse conditions have been documented in Chinese medical literature, ranging from single pulse condition descriptors like floating, rapid, and string-like, each highlighting a specific pulse characteristic, to compound pulse conditions that encompass multiple attributes. For instance, the replete pulse condition is a composite of forceful, long, large, and stiff qualities [2].
2.1. Description of Pulse Condition in Ancient Chinese Medical Texts
The foundational text, Nei Jing [3], details over 30 pulse types, including large, small, long, short, slippery, rough, sunken, slow, rapid, strong, tough, soft, moderate, hurried, vacuous, replete, scattered, intermittent, fine, and weak. Similarly, Mai Jing [4] documents 24 pulse conditions: floating, sunken, hollow, large, small, skipping, tight, rapid, stirred, slippery, weak, string-like, faint, soft, dissipated, moderate, slow, bound, drumskin, replete, intermittent, vacuous, rough, and hidden. The 28 pulse conditions most frequently employed in contemporary clinical practice are derived from Bin Hu Mai Xue [5] and Zhen Jia Zhen Gyan [6], encompassing floating, sunken, slow, rapid, surging, fine, vacuous, replete, long, short, slippery, rough, string-like, tight, soggy, moderate, faint, weak, dissipated, hollow, drumskin, firm, hidden, stirred, intermittent, bound, skipping, and racing.
Ancient Chinese medical texts predominantly describe pulse conditions in qualitative terms, often employing vivid similes and poetic metaphors. For example, the slippery pulse is likened to “beads rolling,” while the string-like pulse is compared to the sensation of pressing a musical instrument string [5]. However, some descriptions, such as rapid, slow, floating, and sunken, possess quantitative aspects. Rapid and slow relate to pulse rate, quantifiable as beats per breath. Floating and sunken describe pulse depth, historically quantified using ‘shu,’ an ancient Chinese unit of weight, with floating corresponding to three shu and sunken to nine shu [7].
The use of analogies and poems to convey pulse conditions introduces inherent subjectivity in interpretation. The string-like pulse, described as feeling like a musical instrument string, and the tight pulse, likened to a rope, rely heavily on the individual TCM doctor’s tactile sensitivity. Qualifying terms such as “a bit,” “average,” and “very” further complicate standardization, as they lack precise quantitative meaning. For example, differentiating between fine and faint pulses based on “a little bit” stronger is inherently imprecise.
Furthermore, overlaps exist between pulse condition descriptions [8,9]. Some conditions describe a single dimension of the pulse, such as floating (depth) or rapid (rate), while others are multi-dimensional. The firm pulse, for instance, encompasses string-like, long, replete, surging, and sunken attributes, and the drumskin pulse is described as string-like, large, rapid, and hollow. The ideal number of dimensions for pulse assessment remains debated. Bin Hu Mai Xue [5] suggests two dimensions: floating/sunken and slow/rapid. In contrast, Nan Jing [7] and Mai Jing [4] propose three: floating/sunken, slippery/rough, and long/short. Nei Jing [3] outlines slippery/rough, slow/rapid, and surging/fine, while other sources [2] suggest floating/sunken, slow/rapid, and vacuous/replete.
The ambiguity in pulse condition descriptions can be attributed to two primary factors. Firstly, TCM doctors traditionally rely on personal perception rather than strictly rationalized criteria for pulse assessment [10]. Secondly, the absence of clear, standardized guidelines for pulse condition diagnosis further exacerbates this issue. These factors likely contribute to the low inter-rater and intra-rater reliability in pulse diagnosis among TCM doctors, as observed in studies by Craddock (1997) and Krass (1990), cited in [11]. In the context of evidence-based practice, which emphasizes outcome consistency [12], the reported low reliability underscores the critical need for standardization in TCM pulse diagnosis.
2.2. Eight Elements: A Milestone in Standardizing TCM Pulse Diagnosis Qualification
Zhou Xuehai (1856-1906) made a significant early stride towards standardizing pulse condition with his proposal that each pulse condition should be defined by four elements. He stated, “Wei Shu Xing Shi Zhe, Zheng Mai Zhi Ti Wang. Qiu Ming Mai Li Zhe, Xu Xian Jiang Wei Shu Xing Shi Jiang De Zhen Qie, Ge Zhong Mai Xiang Liao Ran, Bu Bi Ju Ni Mai Ming” (cited in [13], p. 31). This translates to explicitly identifying position, frequency, shape, and trend as the core elements of any pulse condition, advocating that every pulse condition description should incorporate these four aspects.
Building on this foundational concept, various scholars have expanded and refined Zhou Xuehai’s initial framework [2,14–18], evolving the four elements into eight: depth, rate, regularity, width, length, smoothness, stiffness, and strength. This refined model posits that each pulse condition is characterized by these eight elements, each manifesting with varying degrees of intensity [2,15,17,18].
Rate, in this context, is defined as the number of pulse beats per breath, while regularity mirrors its modern medical definition, describing the rhythm of the pulse. Rate and regularity collectively provide insights into the nature of a disease, particularly differentiating between heat and cold conditions [19]. Depth refers to the vertical position of the pulse, indicating the disease’s location, whether internal or external [19]. Width and length describe the pulse’s spatial characteristics, with width denoting the pulsation intensity and length representing the range over which the pulsation is palpable across the cun, guan, and chi locations [2]. Smoothness describes the pulse’s slickness, stiffness reflects arterial elasticity, and strength indicates the change in pulse forcefulness in response to varying applied pressure [19]. Width, length, smoothness, stiffness, and strength are also considered indicators of the interaction between pathogens and healthy qi within the body [2]. Thus, these eight elements furnish a structured basis for qualifying pulse conditions in TCM, moving towards greater standardization and objectivity.
2.3. Recent Works on Qualifying TCM Pulse Diagnosis
King et al. (2002) [11] developed a measurement scale aiming to standardize TCM pulse diagnosis. However, this scale presents certain limitations in adequately reflecting pulse conditions. Firstly, the six items included—depth, width, force, relative force, rhythm, and pulse occlusion—are not universally recognized as core components of TCM pulse diagnosis. A comprehensive TCM pulse diagnosis necessitates considering the eight elements across all six pulse locations. Secondly, the definitions of these items are somewhat abstract. For instance, “force” is defined as the overall pulse intensity, and “relative force” as a subtler version of overall force. Thirdly, the scale employs an ordinal structure with descriptors for measurement levels, such as depth categorized into superficial, middle, and deep. Ordinal scales, however, may lack the sensitivity required for nuanced pulse assessment due to an insufficient number of response categories [20], and the descriptive terms for each level are not consistently interpreted. Furthermore, given the limited quantification of these items, an ordinal scale may not accurately capture the subtle sensations perceived by a TCM doctor.
To investigate the distinctiveness of each of the eight elements in TCM pulse diagnosis, Tang (2010) [21] employed principal component analysis. The findings revealed that rate, regularity, width, and smoothness represent four unique dimensions, while depth, length, stiffness, and strength did not demonstrate the same level of distinctiveness.
It is crucial to select a content and rating scale that is both appropriate for measuring pulse conditions and aligned with the fundamental concepts of pulse diagnosis in TCM. Given the limited research on qualifying pulse conditions, utilizing a rating scale that genuinely reflects the pulse sensations perceived by TCM doctors is paramount. This approach minimizes the influence of subjective judgment, particularly in the preliminary stages of qualification, paving the way for more objective and reliable assessments.
3. Quantification of TCM Pulse Diagnosis
In the qualification of pulse conditions, the eight elements are assessed unidimensionally. A central hypothesis posits that these eight elements are intrinsically linked to the arterial pressure waveform, and their perceived intensity is a composite of the arterial pressure waveform’s physical parameters. Quantifying these parameters would, therefore, render the eight elements measurable in quantitative terms. Extensive research has been dedicated to the quantification of pulse conditions, primarily employing two main approaches: time domain and frequency domain analysis of the arterial pressure waveform. However, variations in research objectives, methodologies, and statistical approaches have resulted in findings that are often incomparable across studies.
3.1. Time Domain
The time domain approach is extensively utilized in cardiovascular research [22] and has gained considerable traction in quantifying pulse conditions [23]. Time domain analysis examines the arterial pressure waveform in relation to time, graphically depicting its changes over time. Figure 2 illustrates a typical arterial pressure waveform in the time domain.
Figure 2.
A typical arterial pressure waveform (Adapted from [2], p.163)
In time domain analysis, researchers extract physical parameters directly from the arterial pressure waveform, such as h1 and h3, and derive new parameters from these. Yoon et al. (2000) [24] proposed three parameters to quantify depth, width, and strength. Depth was measured by the hold-down pressure corresponding to the maximum h1 (Pamax). Width was quantified using the maximum average h1 (h1), and strength was assessed by the pressure difference at 80% of the maximum average h1 (Δ80%pamax). These parameters have gained some acceptance as standard measures for depth, width, and strength in pulse diagnosis quantification [2,14].
The primary advantage of employing the time domain for quantifying pulse conditions lies in the physiological relevance of most extracted physical parameters. Investigating their relationship with the eight elements holds the potential to elucidate these elements from a modern medical perspective, bridging traditional TCM understanding with contemporary physiological knowledge.
Numerous studies have demonstrated associations between physical parameters of the arterial pressure waveform in the time domain and the eight elements of pulse diagnosis [2,13,24–31]. Depth has been linked to pamax, rate to t (time period), and regularity to the interval consistency between waveforms and their contour uniformity. Width has shown associations with h4/h1, t1, and h1. The surging pulse is characterized by a smaller h4/h1 and t1, and a larger h1. Length has been correlated with h1 at the cun, guan, and chi locations, with the short pulse exhibiting a small h1, although other parameter associations were less distinct. Smoothness has been related to W/t, h4/h1, t1, h5, and h5/h1, with the slippery pulse showing a smaller h4/h1 and a larger h5. Stiffness is associated with h3/h1, h4/h1, and h5/h1, and the string-like pulse exhibits a larger h3/h1 and h4/h1, and a smaller h5/h1. Four waveform types have been identified for the string-like pulse: h1 lower than h3, h3 equal to h1, h3 higher than h1, and h3 merged with h1. Strength is associated with Δ80% pamax. Some of these observations have been interpreted hemodynamically. For example, the string-like pulse is suggested to be caused by increased arterial stiffness and peripheral resistance, while width is influenced by blood velocity, cardiac output, peripheral resistance, radial artery diameter, and radial artery spatial movement. Length is related to the rate of arterial dilatation.
However, inconsistencies across study findings hinder definitive conclusions. Fei (2003) [2] reported depth levels for superficial and deep palpation ranging from 25-175g and 100-250g, respectively, while Xu et al. (2003) [27] reported smaller ranges: less than 100g for superficial, 100-200g for middle, and greater than 200g for deep. Depth is reported in units of force in some studies, while others use pressure units [25,26]. Huang and Sun (1995) [26] reported depth levels of 10-40 mmHg (superficial), 50-80 mmHg (middle), and 90-120 mmHg (deep), whereas Chen (2008) [25] reported considerably higher ranges: 89.8-157.7 mmHg, 151.9-222.9 mmHg, and 279.3 mmHg for superficial, middle, and deep levels, respectively. Regarding smoothness, Huang and Sun (1995) [26] characterized the slippery pulse by t1 in the 0.07-0.09s range, h5 greater than 2 mm, evident h3, and h4/h1 less than 0.50. In contrast, Fei (2003) [2] characterized it by W/t less than 0.20, h4/h1 less than 0.40, and h5/h1 greater than 0.10.
These inconsistencies may stem from several factors. Firstly, none of the studies consistently report the sensor surface area used for pulse acquisition. As force measurements vary with sensor surface area under identical hold-down pressure, this lack of standardization makes cross-study comparisons challenging. Secondly, subject characteristics may influence results. Age, gender, and weight are known factors affecting pulse conditions [2,26], yet demographic data is often lacking in these studies, making it difficult to rule out subject diversity as a source of incongruence. Thirdly, a standardized protocol for pulse waveform acquisition is absent, and procedural details are often scarce in study reports. Mimicking TCM pulse assessment involves applying varying hold-down pressures to the radial artery. Two acquisition procedures are noted. Huang (2007) [13] developed a formula to calculate hold-down pressure for superficial, middle, deep, and hidden depth levels in men and women, based on the ratio of actual to ideal body weight (Table 1).
| (Actual weight)/(Ideal weight) | Hold-down Pressure at Different Levels of Depth |
|---|---|
| Superficial | Middle |
| 50g | 100g |
| 0.8 – 1.0 | 70g |
| 1.0 – 1.2 | 100g |
| “/> 1.2 | 150g |
Table 1.
Weight ratio and corresponding hold-down pressure
While some studies have adopted this protocol [28,32,33], the rationale for quantifying depth in this manner lacks clear explication, raising questions about its validity. The alternative procedure, employed by Fei (2003) [2], involves applying pressure from 0 to 250g in 50g increments for each pulse acquisition. However, the 50g interval may be too coarse to capture subtle waveform changes within this range.
Fourthly, standardized measurements for the eight elements are lacking. Most studies focus on pulse conditions rather than the constituent eight elements. As each pulse condition comprises all eight elements with varying intensities, even with identical intensities across seven elements, variations in a single element can lead to waveform differences for the same pulse condition. Furthermore, the subjective perception of the eight elements varies among TCM doctors, inevitably contributing to result discrepancies across studies.
3.2. Frequency Domain
Frequency domain analysis offers an alternative approach to pulse condition analysis, focusing on the energy distribution within the arterial pressure waveform [34]. A frequency domain graph, derived from the time domain waveform using mathematical transforms like Fast Fourier Transform, consists of amplitude versus frequency and phase versus frequency components. Pulse condition studies typically examine the amplitude versus frequency graph, known as a power spectrum (Figure 3).
Figure 3.
Example of a power spectrum (Adapted from [34])
In a power spectrum, the y-axis (amplitude) represents the power of each frequency component, while the x-axis shows the frequency in Hertz (Hz). A harmonic is a distinct frequency component of the arterial pressure waveform.
Frequency domain analysis in arterial pulse research has primarily focused on disease differentiation [35–37], power spectrum analysis in relation to meridians [38,39], and investigations into disease, syndrome, and channel relationships [23,40–43,45]. Fewer studies have explored power spectrum characteristics specific to different pulse conditions [46,47].
Wang and Xiang (1998) [46] observed marked power spectrum differences between normal, slippery, string-like, and slow-intermittent pulses. Generally, power spectra for all pulse conditions decreased with increasing frequency within the 0-40Hz range. However, the normal pulse exhibited a smoother power spectrum compared to the others. The slippery pulse showed over ten harmonics, while the normal pulse had eight. String-like and slow-intermittent pulses had three to five harmonics. The normal pulse’s frequency was distributed within the 25Hz range. Energy distribution below 10Hz was 99% for normal and 97% for string-like pulses; below 5Hz, it was 90.2% for moderate, 83.7% for slippery, and 60.9% for string-like pulses. For moderate pulses, 45% of energy was below 1 Hz, compared to 16% for string-like pulses. These findings suggest that the normal pulse frequency range is 1-5Hz, with frequencies below 1 Hz and above 10 Hz potentially indicating pathology. Xu et al. (2002) [47] proposed that harmonic count in the power spectrum could differentiate pulse conditions. Their study reported three main harmonics for slippery pulses, significantly higher than normal pulses, and two main harmonics for drumskin pulses. In normal pulses, harmonic amplitude decreased with increasing frequency. The discrepancies in slippery pulse findings, compared to Wang and Xiang (1998), may stem from similar factors as those affecting time domain quantification results.
Both time and frequency domain analyses rely on the arterial pressure waveform but differ in interpretation. While current evidence supporting frequency domain application in pulse condition quantification is less extensive than for time domain, this may simply reflect fewer studies in the frequency domain. Time domain analysis offers a certain advantage due to the physiological interpretability of its parameters, potentially revealing the physiological basis of the eight elements through their relationships with these parameters. Consequently, focusing on the time domain approach may be more prudent and beneficial for advancing the quantification of pulse conditions in chinese pulse diagnosis.
4. Statistical Approaches
While regression analysis is commonly used in medical research for function approximation and classification [48,49], its failure in effectively modeling the relationship between the eight elements and physical parameters [21] suggests that this relationship is non-linear. Advanced statistical techniques, such as fuzzy inference and artificial neural networks (ANN), are considered potentially more suitable for modeling the complex interplay between the eight elements at the six locations and the physical parameters [50–52].
Fuzzy inference is a modeling technique grounded in fuzzy set theory, which addresses the degree of truth within vaguely defined sets, represented on a 0-1 scale. Lee et al. (1993) [53] utilized fuzzy inference to assess the health status of renal patients pre- and post-herbal medicine treatment. Arterial pressure waveforms were acquired at the right chi location, and time domain parameters were used to construct the fuzzy model. The model successfully predicted patient prognosis, leading the authors to propose fuzzy inference for health status assessment via pulse condition analysis.
ANN is a non-linear statistical modeling technique adept at capturing complex non-linear relationships between independent and dependent variables [49,54]. Resembling regression analysis but offering greater flexibility, ANN is unconstrained by statistical assumptions or pre-defined algorithms, making it a self-adaptive and data-driven modeling approach [54,55]. Hidden layers within the network significantly enhance its capacity to handle intricate relationships. Figure 4 illustrates a basic ANN architecture.
Figure 4.
Basic architecture of an ANN (Adapted from [56])
The architecture depicted in Figure 4 is typical of a multilayer perceptron ANN, comprising an input layer, a hidden layer, and an output layer. While input and output layers are also present in linear regression, the hidden layer distinguishes ANN. Researchers can adjust the number of hidden layers to optimize results. The input layer contains input neurons representing independent variables, and output neurons in the output layer represent dependent variables. The number of hidden neurons and layers is determined through trial and error, minimizing the sum-squared error in function approximation and the cross-entropy function in classification. The cross-entropy function serves as a cost function, analogous to the likelihood function in logistic regression [54], guiding model training cessation.
Backpropagation is the most prevalent ANN training algorithm, employing steepest gradient descent in a multilayer perceptron to minimize sum-squared error. Steepest gradient descent is a mathematical algorithm identifying a function’s local minimum by iterative steps proportional to the negative gradient at the current point. In backpropagation, hidden and input neuron weights are adjusted based on sum-squared error feedback from output neurons until mean squared error is minimized.
Wang and Xiang (2001) [57] compared fuzzy inference and ANN accuracy in pulse condition prediction, demonstrating ANN’s successful application in identifying normal, string-like, slippery, and fine pulses, achieving 87% predictive accuracy—12% higher than fuzzy inference. Xu et al. (2007) [58] compared traditional ANN and fuzzy neural network predictive accuracy for eight pulse conditions. Three traditional backpropagation ANNs were developed, each with input, hidden, and output layers. Input neurons represented seventeen time domain physical parameters, and output neurons represented eight unspecified pulse conditions. Hidden neuron counts were 10, 15, and 20 across the three ANNs. The fuzzy neural network, a composite of four sub-fuzzy neural networks, modeled seventeen physical parameters and Zhou Xuehai’s four elements (position, frequency, shape, and trend) separately [13]. These sub-networks were combined to predict the eight pulse conditions. Traditional ANNs achieved 86-88% accuracy, while the fuzzy neural network outperformed them by 4%, suggesting the benefit of combining fuzzy inference and ANN in quantifying pulse conditions.
The successful application of advanced statistical techniques in quantifying pulse conditions is encouraging, indicating a physiological basis for various pulse conditions. However, medical research prioritizes model explanatory power [59–61], favoring techniques with high interpretability. ANN, in this context, can be criticized as a “black box” [49], meaning its internal workings are not readily accessible to researchers [59], limiting the understanding of underlying mechanisms despite predictive success.
5. TCM Pulse Diagnostic Framework: Integrating Eastern and Western Perspectives
A dice model, formulated by Tang (2010) [21], offers a framework to explain the interconnectedness between arterial pulse and the eight elements of pulse condition across the six locations, and subsequently between these elements and overall health status in TCM. This model serves as a foundational structure for quantifying chinese pulse diagnosis, bridging traditional Eastern concepts with Western scientific approaches.
The dice model operates on two levels. Level one encompasses the arterial pulse and the eight elements at the six locations, focusing on the sensation of arterial pulse as perceived by a TCM doctor. Level two connects the eight elements at the six locations to health status in TCM, interpreting these elements to determine overall well-being. These levels are intrinsically interconnected. The symbolic representation of a dice and a dice roll within this model elucidates the relationship between arterial pulse and health status in TCM.
5.1. Arterial Pulse and the Eight Elements at the Six Locations
The model posits that the eight elements—depth, rate, regularity, width, length, smoothness, stiffness, and strength—are influenced by the arterial pulse at the six locations (left and right cun, guan, and chi). The intensity of each element is determined by the TCM doctor’s perception of the arterial pulse. Consequently, the eight elements at the six locations are operationalized as ratings along a continuum, with Yin and Yang representing opposing extremes.
Specifically, depth is operationalized as the vertical position of the arterial pulse, rated from deepest (Yin) to most floating (Yang). Rate is quantified as beats per minute, ranging from slowest (Yin) to most rapid (Yang). Regularity is categorized as either regular or irregular. Width represents pulse intensity, from smallest (Yin) to largest (Yang). Length is the range of the pulse palpable across cun, guan, and chi, from shortest (Yin) to longest (Yang). Smoothness is the pulse’s slickness, from roughest (Yin) to smoothest (Yang). Stiffness reflects radial artery elasticity, from least stiff (Yin) to stiffest (Yang). Finally, strength is the pulse forcefulness relative to applied pressure changes, from least forceful (Yin) to most forceful (Yang).
5.2. The Eight Elements at the Six Locations and Health Status
In TCM pulse diagnosis, health status is assessed based on pulse conditions at the six locations, each reflecting the health of specific organs. Left cun, guan, and chi correspond to the heart, liver, and kidneys, respectively, while right cun, guan, and chi represent the lungs, spleen, and kidneys (lifegate). The eight elements serve as criteria for evaluating organ health. Overall health status, the outcome measure of TCM pulse diagnosis, is a composite assessment of individual organ health.
5.3. The Dice Model
The dice model visually represents the intertwined and cascading relationships between arterial pulse, the eight elements at the six locations, and health status (Figure 5).
Figure 5.
The dice model
5.3.1. Assumptions
The dice model is built upon three key assumptions: firstly, that all eight elements are equally weighted in overall pulse condition assessment; secondly, that the midpoint of each element’s continuum signifies Yin-Yang balance; and thirdly, that all six locations are equally important in determining overall health status.
5.3.2. Symbolic Meaning
The dice itself symbolizes the TCM concept of health, viewed as a balance between Yin and Yang, which in turn relies on the individual function and interaction of various organs. The six pyramids forming the dice represent the organs at the six pulse locations.
The interior of the dice represents blood flow within the organs, the collective of which constitutes the arterial pulse. Changes in organ blood flow are reflected in the arterial pulse. Assessing the six pyramids reveals the health status of individual organs and, consequently, overall health.
5.3.3. Position of the Six Organs
The six pyramids represent the six pulse locations. Lung and heart, liver and spleen, and kidneys and lifegate are positioned in opposing pyramids, reflecting their roles in overall health. This arrangement aligns with the notion that left cun, guan, and chi assess blood (Yin), while right cun, guan, and chi assess qi (Yang), adhering to Yin-Yang theory.
5.3.4. The Eight Elements
Each pyramid is composed of the eight elements, with the enlarged square in Figure 5 illustrating their interrelation. Each element is a complementary Yin-Yang pair. In line with Yin-Yang theory, Yin (black square) represents the internal aspect, forming the pyramid’s core, and Yang (white square) represents the external aspect, forming the outer part.
The intensity of the eight elements is determined by the arterial pulse. The combined intensity of these elements within each pyramid reflects the health status of the corresponding organ.
5.3.5. Interconnection in the Dice Model
The dice model is inspired by the Taiji symbol. The dotted lines connecting the six pyramids symbolize the dynamic and interchanging relationships among organs. Dotted lines also connect Yin and Yang within each element, and the eight elements within each pyramid, signifying their Yin-Yang composite nature and constant interchange and balancing. The solid dice outline represents the absolute state of health, akin to the Taiji circle representing the universe. Health is considered neither expandable nor reducible; only health status fluctuates, determined by Yin-Yang interaction within the body.
5.3.6. Analogy Between a Dice Roll and Health Status
A dice roll serves as an analogy for health status. A “fair” dice, with equal pyramid areas and weights, has an equal probability of landing on each pyramid, representing a healthy state. In this state, organ blood flow is normal, wave reflection and resonance are proper, eight element intensities are near the continuum midpoint forming a regular pyramid shape, and the six pyramids are balanced, reflecting Yin-Yang harmony.
A “loaded” dice, with altered pyramid areas or weights, represents an unhealthy state. Unequal pyramid probabilities in a loaded dice mirror unhealthy status, where organ dysfunction alters circulatory wave reflection and resonance, changing blood flow and pyramid weights. Arterial pulse and eight element intensities shift accordingly. The pyramid shape becomes irregular, skewed, smaller, or larger, leading to imbalance among the six pyramids and Yin-Yang disharmony, compromising health.
5.4. Recent Works on Validating the Framework
Several studies have aimed to validate the dice model framework. Tang et al. (2012) [62] conducted a study to validate the content and diagnostic accuracy, achieving a content validity index of 0.73, considered acceptable. Criterion validation, using artificial neural networks, yielded approximately 80% accuracy across models, with sensitivity and specificity ranging from 70% to nearly 90%, supporting the framework’s content and diagnostic validity.
Tang et al. (2012) [63] reported successful establishment of non-linear relationships between the eight elements at six locations and time domain physical parameters using the Levenberg-Marquardt algorithm, with r-squared values ranging from 0.60 to 0.86, further supporting the framework’s underlying principles.
6. Conclusion: Charting the Future of TCM Pulse Diagnosis Quantification
The TCM pulse diagnostic framework, proposed by Tang (2010) [21], offers a promising direction for guiding the quantification of chinese pulse diagnosis. While initial studies [62,63] have provided preliminary verification of its hypotheses, further research is essential for comprehensive validation. This section outlines current study limitations and recommends future research directions.
6.1. Limitations
6.1.1. Methodology
TCM doctors typically assess pulse at all six locations simultaneously and individually. However, current pulse acquisition devices are predominantly single-probe, limiting waveform acquisition to one location at a time due to the lack of validated three-sensor devices. This constraint prevents the examination of simultaneous pulse manipulation across all six locations, a key aspect of traditional TCM pulse diagnosis.
The pulse acquisition process, lasting approximately one hour in studies [62], [63], may induce motion artifacts due to subject movement, impacting waveform quality. Baseline fluctuations in waveforms, caused by subject movement and breathing, were not adequately addressed by feature extraction programs. Rescaling fluctuating baselines to a horizontal level could distort waveforms and introduce feature extraction errors. Furthermore, the peak-to-peak interval parameter could not be extracted by the program, limiting its use in modeling. Device limitations in hold-down pressure (max 400 mmHg) prevented Δ80%pamax calculation in some subjects, as waveform amplitude did not consistently decrease.
Subject characteristics also present a limitation. Studies [62], [63] recruited relatively stable subjects, resulting in a narrow intensity range for the eight elements. This homogeneity may have reduced r-squared values in statistical models, limiting the generalizability of findings.
6.1.2. Statistical Approaches
ANNs were chosen to analyze framework relationships due to their non-linear nature. However, ANN application faces limitations. Large sample sizes, typically required for robust ANN modeling, are challenging to achieve in clinical studies. Smaller sample sizes may have reduced model effect sizes. As previously mentioned, ANN’s limited explanatory power hinders in-depth model analysis, despite their predictive capabilities, limiting mechanistic understanding of chinese pulse diagnosis.
6.2. Recommendations
Addressing the identified limitations, five recommendations are proposed for future research to advance the quantification of chinese pulse diagnosis.
6.2.1. Methodology
Developing a validated three-sensor pulse acquisition device is crucial to enable the study of simultaneous hold-down pressure effects on arterial pressure waveforms at cun, guan, and chi. Enhancing feature extraction programs is also necessary to extract all relevant parameters from waveforms, as these parameters are fundamental for physical parameter calculation and subsequent analysis.
6.2.2. Statistical Approaches
Developing programs capable of elucidating underlying relationships between physical parameters and the eight elements at the six locations, and between these elements and health status, is recommended. Enhancing model explanatory power will provide stronger scientific backing for TCM pulse diagnosis, bolstering evidence for TCM theories and improving the understanding of chinese pulse diagnosis.
Further recommendations include diversifying subject recruitment to validate models across broader populations. Current models, based on stable hypertensive subjects, are preliminary and may not generalize to severe hypertension or hypotension. Future studies should recruit subjects with varying degrees of hypertension and hypotension to improve model generalizability. Additionally, including patients with other diseases will enable the assessment of the models’ ability to differentiate hypertension from other disease states, enhancing the clinical utility and specificity of chinese pulse diagnosis quantification.
