Pulse Diagnosis in Traditional Chinese Medicine (TCM): An Overview of Quantification and Research

1. Introduction to TCM Pulse Diagnosis

Traditional Chinese Medicine (TCM) pulse diagnosis stands as a cornerstone of TCM consultation, recognized as one of the four essential diagnostic methods. This technique involves pulse palpation at three specific locations – Cun, Guan, and Chi – on both wrists, offering a comprehensive assessment of an individual’s overall health and the condition of particular organs. Figure 1 illustrates these locations and their corresponding organ associations, a critical map in the practice of TCM pulse diagnosis. TCM practitioners integrate the insights from pulse assessment with other clinical data to formulate personalized treatments and monitor patient progress.

Figure 1. Organ Distribution in TCM Pulse Diagnosis Locations

Note: This image depicts the traditional organ associations for each pulse location in TCM pulse diagnosis.

The increasing global interest in TCM has brought pulse diagnosis into the spotlight, sparking public curiosity about its scientific validity and clinical effectiveness. Since the 1950s, extensive research has been dedicated to quantifying TCM pulse diagnosis, aiming to establish a scientific foundation and validate its clinical utility. This article provides a thorough overview of the current advancements in the quantification of TCM pulse diagnosis, offering readers a comprehensive understanding of this evolving field.

Upon reading this article, you will:

1. Gain insight into the latest scientific evidence supporting the quantification of TCM pulse diagnosis.
2. Develop a critical perspective on the methodologies and statistical approaches used in current research, identifying both strengths and weaknesses.
3. Understand the future directions in the scientific study of TCM pulse diagnosis quantification.

This review is structured into five sections, beginning with an exploration of the qualification and quantification of TCM pulse diagnosis in both historical and contemporary literature. We will discuss statistical methodologies applied to quantify TCM pulse diagnosis and, in Section Four, introduce a TCM pulse diagnostic framework proposed in 2010, designed to clarify the relationship between pulse conditions and arterial pulse. Finally, the concluding section addresses the limitations of existing research and suggests recommendations for future investigations.

2. Qualifying TCM Pulse Diagnosis: Defining the Elements

The qualification of TCM pulse diagnosis refers to defining the essential elements necessary for a complete and accurate pulse assessment. Literature review reveals considerable ambiguity in pulse assessment within TCM, primarily due to the often-vague descriptions of pulse conditions 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 quality of the pulse as perceived and interpreted by a TCM doctor, reflecting their professional judgment. Over 30 distinct pulse conditions have been documented in Chinese medical literature. Some are single pulse conditions, describing a singular aspect, such as “floating,” “rapid,” or “string-like.” Others are compound pulse conditions, integrating multiple aspects, for example, “replete,” which combines “forceful,” “long,” “large,” and “stiff” characteristics [2].

2.1. Historical Descriptions of Pulse Conditions in Chinese Medicine

The Nei Jing (The Yellow Emperor’s Classic of Internal Medicine) [3] documents 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.” The Mai Jing (Pulse Classic) [4] details 24 types: “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 commonly used in contemporary clinical practice are derived from Bin Hu Mai Xue (Bin Hu’s Study of the Pulse) [5] and Zhen Jia Zhen Gyan (Pearls of Zhen Jia) [6]. These include “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.”

Descriptions of pulse conditions in ancient Chinese medical texts are predominantly qualitative, often employing analogies and poetic language. For example, a “slippery” pulse is likened to “beads rolling,” while a “string-like” pulse feels like “pressing the string of a musical instrument” [5]. However, some descriptions, such as “rapid,” “slow,” “floating,” and “sunken,” offer a degree of quantification. “Rapid” and “slow” refer to pulse rate, quantifiable by beats per breath. “Floating” and “sunken” describe pulse depth, quantified using the ancient Chinese unit of weight, shu, with “floating” corresponding to three shu and “sunken” to nine shu [7].

The use of analogies and poems to describe pulse conditions introduces subjectivity, dependent on the TCM doctor’s interpretation. For instance, the sensation of “string-like” or “tight” depends on the individual practitioner’s tactile sensitivity. Qualifying terms like “a bit,” “average,” and “very” are used to denote intensity, but lack precise, objective measurement. The subtle difference between “fine” and “faint,” for example, described as “a little bit stronger,” remains ambiguous in its quantifiable distinction.

Furthermore, overlaps exist between pulse condition descriptions [8,9]. Some conditions describe a single dimension, like “floating” (depth) or “rapid” (rate). Others encompass multiple dimensions. “Firm” integrates “string-like,” “long,” “replete,” “surging,” and “sunken,” while “drumskin” combines “string-like,” “large,” “rapid,” and “hollow.” The number of essential dimensions for pulse assessment is debated. Bin Hu Mai Xue [5] suggests “floating/sunken” and “slow/rapid.” Nan Jing (Classic of Difficult Issues) [7] and Mai Jing [4] propose three: “floating/sunken,” “slippery/rough,” and “long/short.” Nei Jing [3] also describes three: “slippery/rough,” “slow/rapid,” and “surging/fine,” while others [2] suggest “floating/sunken,” “slow/rapid,” and “vacuous/replete.”

Two primary reasons contribute to the ambiguity in pulse condition descriptions. First, TCM practitioners often rely on personal perception rather than strictly defined rational criteria [10]. Second, a lack of concise and precise standards to guide pulse diagnosis exists. These factors likely contribute to the low inter-rater and intra-rater reliability of pulse diagnosis among TCM doctors, as noted by Craddock (1997) and Krass (1990), cited in [11]. In evidence-based practice, consistency is paramount [12]. The reported low reliability highlights the critical need for standardization in TCM pulse diagnosis.

2.2. The Eight Elements: A Standardizing Milestone

Zhou Xuehai (1856-1906) made an early and significant attempt to standardize pulse conditions, a landmark in the quantification of TCM pulse diagnosis. He proposed that each pulse condition should be defined by four elements: “position, frequency, shape, and trend.” He asserted that a clear understanding of these elements – position, frequency, shape, and trend – is essential for accurate pulse assessment, enabling a comprehensive understanding of various pulse manifestations without being constrained by pulse names (cited in [13], p. 31).

Building on this foundation, scholars have expanded these initial four elements into eight [2,14–18]: depth, rate, regularity, width, length, smoothness, stiffness, and strength. The consensus is that every pulse condition is characterized by these eight elements, each with varying intensities [2,15,17,18].

• Rate is defined as the number of pulse beats per breath, providing insights into the nature of disease, whether heat or cold [19].
• Regularity mirrors the modern medical definition, describing the pulse rhythm, also indicating disease nature [19].
• Depth refers to the vertical position of the pulse, indicating the disease’s location, whether interior or exterior [19].
• Width and length describe the pulse’s spatial characteristics. Width is the pulsation intensity, and length is the range over which pulsation is sensed across Cun, Guan, and Chi [2].
• Smoothness is the pulse’s slickness, while stiffness is the sensation of arterial elasticity.
• Strength is the pulse’s forcefulness in response to varying applied pressure [19].

Width, length, smoothness, stiffness, and strength also reflect the interaction between pathogens and the body’s healthy Qi [2]. Collectively, these eight elements provide a structured framework for qualifying pulse conditions.

2.3. Contemporary Approaches to Qualifying TCM Pulse Diagnosis

King et al. (2002) [11] developed a measurement scale aimed at standardizing TCM pulse diagnosis. However, this scale has limitations in adequately representing pulse conditions. Firstly, the six items it includes – depth, width, force, relative force, rhythm, and pulse occlusion – are not universally recognized as core components in TCM pulse diagnosis. A comprehensive scale should incorporate the eight elements across the six pulse locations. Secondly, the definitions of these items are abstract. For instance, “force” is defined as overall pulse intensity, and “relative force” as a subtler version. Thirdly, the scale uses an ordinal scale with descriptors to measure these items. Depth, for example, is measured at three levels: superficial, middle, and deep. Ordinal scales lack sensitivity due to limited response categories [20], and the descriptive terms for each level are not consistently interpreted. Given the lack of quantification for these items, an ordinal scale may not accurately reflect the sensations perceived by TCM practitioners.

To investigate the uniqueness of each of the eight elements, Tang (2010) [21] employed principal component analysis. The findings indicated that rate, regularity, width, and smoothness represent four distinct dimensions, while depth, length, stiffness, and strength did not demonstrate the same level of uniqueness.

The author argues that selecting an appropriate content and rating scale is crucial for accurately measuring pulse conditions. This scale must be relevant to and consistent with the fundamental principles of TCM pulse diagnosis. Given the limited research in qualifying pulse conditions, a rating scale that genuinely captures the pulse sensations perceived by TCM doctors is essential, particularly in the preliminary stages of qualification, to minimize subjective bias.

3. Quantification of TCM Pulse Diagnosis: Measuring the Immeasurable

In the qualification process, the eight elements of pulse diagnosis are considered unidimensionally. It is theorized that these elements are linked to the arterial pressure waveform, and their perceived intensity is a composite of the waveform’s physical parameters. Quantifying these parameters could provide a measurable basis for the eight elements. Extensive research has focused on quantifying pulse conditions, primarily using time domain and frequency domain analyses of the arterial pressure waveform. However, variations in research objectives, methodologies, and statistical approaches have resulted in findings that are often difficult to compare directly.

3.1. Time Domain Analysis

Time domain analysis, a widely used method in cardiovascular research [22], is also prevalent in quantifying pulse conditions [23]. This approach examines the arterial pressure waveform over time, depicting how it changes. Figure 2 illustrates a typical arterial pressure waveform in the time domain.

Figure 2. Typical Arterial Pressure Waveform in Time Domain

Note: This figure represents a standard arterial pressure waveform, illustrating key parameters used in time domain analysis for TCM pulse diagnosis quantification.

Researchers extract physical parameters from the arterial pressure waveform in the time domain, such as h1, 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 with the largest h1 (Pamax), width by the maximum average h1 (h1), and strength by the pressure difference at 80% maximum average h1 (Δ80%pamax). These parameters are increasingly recognized as standard measures for depth, width, and strength [2,14].

The advantage of time domain analysis is that most related physical parameters have physiological significance. Exploring their relationship with the eight elements can enhance our understanding of these elements from a modern medical perspective.

Numerous studies have demonstrated associations between time domain parameters and the eight elements [2,13,24–31]. Depth correlates with pamax, rate with t, and regularity with the interval and contour consistency of waveforms. Width is linked to h4/h1, t1, and h1. “Surging” pulses show smaller h4/h1 and t1, and larger h1. Length relates to h1 at Cun, Guan, and Chi; “short” pulses exhibit small h1. Smoothness is associated with W/t, h4/h1, t1, h5, and h5/h1, with “slippery” pulses showing smaller h4/h1 and larger h5. Stiffness is linked to h3/h1, h4/h1, and h5/h1; “string-like” pulses show larger h3/h1 and h4/h1, and smaller h5/h1, with waveform variations including h1<h3, h3=h1, h3>h1, and h3 merged with h1. Strength correlates with Δ80% pamax. Hemodynamic explanations exist for some observations. For example, “string-like” pulses are linked to increased arterial stiffness and peripheral resistance, while width is influenced by blood velocity, cardiac output, peripheral resistance, radial artery diameter, and its spatial movement. Length relates to arterial dilatation rate.

However, inconsistencies across studies hinder definitive conclusions. Fei (2003) [2] reported depth levels ranging from 25-175g (superficial) to 100-250g (deep), while Xu et al. (2003) [27] reported smaller ranges: <100g (superficial), 100-200g (middle), and >200g (deep). Depth is reported in units of force in some studies, and pressure in others [25,26]. Huang and Sun (1995) [26] reported depth ranges of 10-40 mmHg (superficial), 50-80 mmHg (middle), and 90-120 mmHg (deep), while Chen (2008) [25] reported 89.8-157.7 mmHg, 151.9-222.9 mmHg, and 279.3 mmHg respectively. For smoothness, Huang and Sun (1995) [26] characterized “slippery” pulses with t1 of 0.07-0.09s, h5>2mm, obvious h3, and h4/h1<0.50, whereas Fei (2003) [2] found “slippery” pulses with W/t<0.20, h4/h1<0.40, and h5/h1>0.10.

Four main factors contribute to these inconsistencies. First, studies often fail to report sensor surface area, making force measurements incomparable. Second, subject characteristics (age, gender, weight) significantly affect pulse conditions [2,26], yet demographic data is often lacking. Third, a standardized pulse acquisition protocol is absent. Mimicking TCM pulse assessment involves applying varying hold-down pressures. Two procedures exist: Huang (2007) [13] developed a formula based on weight ratio to calculate hold-down pressure for different depth levels (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 Hold-Down Pressure in Pulse Acquisition

Note: This table details the protocol proposed by Huang (2007) for adjusting hold-down pressure in TCM pulse diagnosis quantification based on the subject’s weight ratio.

While some studies adopt this protocol [28,32,33], its rationale for quantifying depth remains unclear. Fei’s (2003) [2] procedure involves pressure application from 0-250g at 50g intervals. However, 50g intervals may be too broad to capture waveform changes effectively. Fourth, a standard measurement for the eight elements is lacking. Most studies focus on pulse conditions rather than the eight elements themselves. As each pulse condition encompasses all eight elements with varying intensities, even if seven elements are consistent, a change in one element can alter the waveform. Furthermore, TCM doctors’ perceptions of the eight elements vary, contributing to result discrepancies.

3.2. Frequency Domain Analysis

Frequency domain analysis offers another approach to pulse condition analysis, focusing on the energy distribution within the arterial pressure waveform [34]. This method converts the time domain waveform into a frequency domain graph, typically examining the amplitude versus frequency component, known as the power spectrum (Figure 3).

Figure 3. Example of a Power Spectrum in Frequency Domain Analysis

Note: This power spectrum illustrates the amplitude of frequency components in an arterial pressure waveform, a key aspect of frequency domain analysis in TCM pulse diagnosis research.

In a power spectrum, the y-axis represents the “amplitude” or power of each frequency, and the x-axis shows “frequency” in Hertz (Hz). A harmonic is a frequency component of the arterial pressure waveform.

Frequency domain analysis in TCM pulse research has primarily focused on disease differentiation [35–37], power spectrum analysis in relation to meridians [38,39], and the interrelation of disease, syndrome, and channels [23,40–43,45]. Fewer studies have explored power spectrum characteristics for different pulse conditions [46,47].

Wang and Xiang (1998) [46] observed distinct power spectra for “normal,” “slippery,” “string-like,” and “slow-intermittent” pulses. Generally, power spectra decreased with increasing frequency within 0-40Hz. The “normal” pulse spectrum was smoother than others. “Slippery” pulses had over ten harmonics, “normal” pulses eight, and “string-like” and “slow-intermittent” pulses three to five. The frequency range for “normal” pulses was within 25Hz. Energy distribution below 10Hz was 99% for “normal” and 97% for “string-like” pulses, while below 5Hz, it was 90.2% for “moderate,” 83.7% for “slippery,” and 60.9% for “string-like.” For “moderate” pulses, 45% of energy was below 1Hz, compared to 16% for “string-like” pulses. These findings suggest “normal” pulse frequencies fall within 1-5Hz, with frequencies below 1Hz and above 10Hz potentially indicating pathology. Xu et al. (2002) [47] proposed harmonic count in power spectra to differentiate pulse conditions. Their study indicated “slippery” pulses had three main harmonics much higher than “normal,” and “drumskin” pulses had two. In “normal” pulses, harmonic amplitude decreased with frequency. Discrepancies in “slippery” pulse results are attributed to similar factors as in time domain quantification.

Both time domain and frequency domain analyses utilize the arterial pressure waveform but interpret it differently. While frequency domain evidence for pulse condition quantification is currently less robust than time domain, this may reflect fewer studies in this area. Time domain analysis may offer advantages due to the physiological relevance of its parameters, potentially revealing physiological implications of the eight elements. Therefore, time domain analysis may be a more promising approach for quantifying pulse conditions.

4. Statistical Approaches in TCM Pulse Diagnosis Research

Regression analysis, commonly used in medical research for function approximation and classification [48,49], has proven inadequate for modeling the relationship between the eight elements and physical parameters [21], suggesting a non-linear relationship. Advanced statistical techniques, such as fuzzy inference and artificial neural networks (ANN), may be more suitable for modeling the complex relationships between the eight elements at the six locations and physical parameters [50–52].

Fuzzy inference is a modeling technique based on fuzzy set theory, which deals with degrees of truth in vaguely defined sets, represented by values from 0 to 1. Lee et al. (1993) [53] used fuzzy inference to assess the health status of renal patients before and after herbal medicine. Using time domain parameters from the right Chi pulse, their fuzzy model successfully predicted patient prognosis, suggesting its potential for health status assessment via pulse condition.

ANN, a non-linear statistical modeling technique, excels in modeling complex non-linear relationships between variables [49,54]. Resembling regression analysis but with greater flexibility, ANN is data-driven and self-adaptive, unconstrained by statistical assumptions [54,55]. Hidden layers enhance its capacity to handle complex relationships. Figure 4 shows a basic ANN architecture.

Figure 4. Basic Artificial Neural Network (ANN) Architecture

Note: This diagram illustrates the fundamental structure of an Artificial Neural Network, commonly used in TCM pulse diagnosis research for complex data analysis and pattern recognition.

The architecture in Figure 4, typical of a multilayer perceptron ANN, includes an input layer, hidden layer, and output layer. Unique to ANN is the hidden layer(s) between input and output. The number of hidden layers and neurons is adjusted to optimize results, assessed by sum-squared error in function approximation and cross-entropy function in classification, analogous to the likelihood function in logistic regression [54]. These cost functions determine model training completion.

Backpropagation, a popular ANN training algorithm, uses steepest gradient descent in multilayer perceptrons to minimize sum-squared error. This algorithm iteratively adjusts hidden and input neuron weights based on error feedback from output neurons until mean squared error is minimized.

Wang and Xiang (2001) [57] compared fuzzy inference and ANN accuracy in predicting pulse conditions. ANN successfully identified “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 accuracy in predicting eight pulse conditions. Three traditional ANNs using backpropagation, each with three layers and 17 time domain parameters as input, were compared against a fuzzy neural network composed of four sub-networks modeling 17 parameters and Zhou Xuehai’s four elements (position, frequency, shape, trend). Traditional ANNs achieved 86-88% accuracy, while the fuzzy neural network outperformed them by 4%, suggesting the benefit of combining fuzzy inference and ANN.

The successful application of these advanced statistical techniques in quantifying pulse conditions is promising, indicating a physiological basis for various pulse conditions. However, medical research values model explanatory power [59–61]. ANN, often considered a “black box” [49], lacks transparency in its internal processes, limiting researchers’ ability to fully understand the models generated [59].

5. TCM Pulse Diagnostic Framework: Bridging Eastern and Western Perspectives

Tang (2010) [21] proposed a “dice model” framework to explain the interconnections between arterial pulse, the eight elements of pulse condition at the six locations, and health status in TCM. This framework provides a structure for quantifying TCM pulse diagnosis.

The dice model operates on two levels. Level one connects arterial pulse and the eight elements at the six locations, while level two links these elements to health status in TCM. Level one addresses the TCM doctor’s sensation of the arterial pulse, and level two interprets the eight elements to determine health status. These levels are interconnected. The dice model uses the analogy of a dice and a dice roll to represent arterial pulse and health status in TCM.

5.1. Arterial Pulse and the Eight Elements

The model posits that the eight elements are influenced by the arterial pulse at the six locations (left and right Cun, Guan, and Chi). The intensity of each element (depth, rate, regularity, width, length, smoothness, stiffness, and strength) is determined by the TCM doctor’s perception of the arterial pulse. Each element is operationalized as a rating on a Yin-Yang continuum.

• Depth: Vertical pulse position, rated from deepest (Yin) to most floating (Yang).
• Rate: Beats per minute, from slowest (Yin) to fastest (Yang).
• Regularity: Pulse rhythm, categorized as regular or irregular.
• Width: Pulse intensity, from smallest (Yin) to largest (Yang).
• Length: Pulse range across Cun, Guan, Chi, from shortest (Yin) to longest (Yang).
• Smoothness: Pulse slickness, from roughest (Yin) to smoothest (Yang).
• Stiffness: Arterial elasticity, from least stiff (Yin) to stiffest (Yang).
• Strength: Pulse forcefulness relative to applied pressure, from least (Yin) to most forceful (Yang).

5.2. Eight Elements and Health Status

In TCM pulse diagnosis, health status is determined by pulse conditions at the six locations, each reflecting the health of a specific organ. Left Cun, Guan, and Chi correspond to the heart, liver, and kidneys; right Cun, Guan, and Chi to the lungs, spleen, and kidneys (life-gate). The eight elements are the assessment criteria for organ health. Overall health status, a composite measure, is derived from the health status of these organs.

5.3. The Dice Model Analogy

The dice model uses a dice to symbolize the complex relationships between arterial pulse, the eight elements, and health status (Figure 5).

5.3.1. Model Assumptions

The dice model is based on three assumptions:

1. All eight elements are equally weighted in pulse condition assessment.
2. The midpoint of each element’s continuum represents Yin-Yang balance.
3. All six locations are equally weighted in determining health status.

5.3.2. Dice Symbolism

The dice represents the TCM concept of health as Yin-Yang balance, reliant on organ function and interaction. The six pyramids forming the dice symbolize the six pulse locations/organs.

The dice’s interior represents organ blood flow, collectively forming the arterial pulse. Changes in organ blood flow are reflected in the arterial pulse. Assessing the six pyramids reveals organ and overall health status.

Figure 5. The Dice Model of TCM Pulse Diagnosis

Note: This image presents the Dice Model framework for TCM pulse diagnosis, illustrating the interconnectedness of arterial pulse, eight elements, and health status, using a dice analogy.

5.3.3. Organ Positioning

The six pyramids represent the six pulse locations. Lung/Heart, Liver/Spleen, and Kidney/Life-gate are positioned on opposite pyramids, reflecting their health roles. Left Cun/Guan/Chi assess blood (Yin), and right Cun/Guan/Chi assess Qi (Yang), aligning with Yin-Yang theory.

5.3.4. The Eight Elements Representation

Each pyramid is composed of the eight elements, depicted in the enlarged square. Each element is a Yin-Yang pair. The black (Yin) square is the pyramid’s core, and the white (Yang) square is the outer part, consistent with Yin-Yang theory where Yin is internal and Yang external.

Element intensity depends on the arterial pulse. The combined intensity of the eight elements indicates the health of the organ represented by that pyramid.

5.3.5. Interconnections within the Model

Inspired by the Taiji symbol, the dice model uses dotted lines to connect the six pyramids, symbolizing dynamic organ relationships. Dotted lines also connect Yin-Yang aspects within elements and elements within pyramids, indicating their Yin-Yang composite nature and constant interplay. The solid dice outline represents absolute health, analogous to the Taiji circle representing the world – health is finite, while health status fluctuates with Yin-Yang interaction.

5.3.6. Dice Roll Analogy to Health Status

A dice roll further illustrates the model. A “fair” dice, with equal pyramid areas and weights, symbolizes health. Equal probability for each pyramid represents balanced organ blood flow, proper wave reflection/resonance, mid-continuum eight-element intensities forming regular pyramid shapes, and balanced pyramids. Yin-Yang harmony is achieved.

An altered “loaded” dice, with uneven pyramid areas or weights, symbolizes unhealthy status. Unequal pyramid probabilities reflect organ dysfunction impacting wave reflection/resonance, altered blood flow, and skewed pyramid weights. Arterial pulse changes reflect organ health, altering eight-element intensities and leading to irregular, skewed pyramids. Yin-Yang imbalance and compromised health ensue.

5.4. Validation Studies of the Dice Model

Several studies have begun to validate the dice model framework. Tang et al. (2012) [62] assessed the framework’s content and diagnostic validity, achieving an acceptable content validation index of 0.73. Criterion validation using artificial neural networks showed approximately 80% accuracy across ANN models, with specificity and sensitivity ranging from 70% to nearly 90%, supporting the framework’s content and diagnostic validity.

Tang et al. (2012) [63] also demonstrated the successful establishment of non-linear relationships between the eight elements and time domain physical parameters using the Levenberg-Marquardt algorithm, with r-squared values ranging from 0.60-0.86.

6. Conclusion: Future Directions in TCM Pulse Diagnosis Quantification

The TCM pulse diagnostic framework proposed by Tang (2010) [21] provides a promising direction for quantifying TCM pulse diagnosis. While initial validation studies [62,63] are encouraging, they are preliminary and require further research. This section outlines current limitations and offers recommendations for future studies in Pulse Diagnosis In Tcm.

6.1. Limitations of Current Research

6.1.1. Methodological Limitations

TCM doctors typically assess pulse at all six locations simultaneously and individually. Current pulse acquisition devices are mostly single-probe, hindering simultaneous measurement due to the lack of validated multi-sensor devices. This limits the study of simultaneous pulse manipulation characteristic of TCM diagnosis.

Prolonged pulse acquisition (around one hour [62], [63]) may induce motion artifacts, affecting waveform quality. Baseline fluctuations due to subject movement and breathing, and inadequate noise removal by current feature extraction programs are also issues. Rescaling fluctuating baselines can distort waveforms and introduce feature extraction errors. Furthermore, peak-to-peak interval extraction program limitations and a device pressure limit of 400 mmHg, preventing ∆80%pamax calculation in some subjects, pose additional challenges.

Subject homogeneity is another limitation. Studies recruiting stable subjects [62], [63] may have narrowed the range of eight-element intensities, potentially lowering r-squared values in statistical analyses.

6.1.2. Statistical Approach Limitations

ANN, chosen for its ability to model non-linear relationships, faces limitations. The large sample sizes ANN requires for robust results are often unattainable in clinical studies, potentially reducing model effect sizes with smaller samples. Additionally, ANN’s “black box” nature limits explanatory power, hindering in-depth model analysis.

6.2. Recommendations for Future Research

Addressing these limitations, five key recommendations are proposed for future research in TCM pulse diagnosis quantification.

6.2.1. Develop Advanced Methodologies

Developing a validated three-sensor pulse acquisition device is crucial to study simultaneous hold-down pressure effects on arterial pressure waveforms at Cun, Guan, and Chi. Enhanced feature extraction programs are needed to accurately capture all necessary waveform features, as physical parameter calculations depend on these features.

6.2.2. Enhance Statistical Approaches and Explanatory Power

Developing programs capable of elucidating underlying relationships between physical parameters and the eight elements, and between the eight elements and health status, is recommended. Improving model explanatory power would provide robust scientific backing for TCM pulse diagnosis and strengthen the evidence supporting TCM theories.

6.2.3. Broaden Subject Recruitment

Future studies should recruit more diverse subjects to validate models across a wider spectrum. Current models, developed with stable hypertensive subjects, cannot be readily generalized to severe hypertension or hypotension cases. Recruiting subjects with varying degrees of hypertension and hypotension is recommended to improve model generalizability. Furthermore, including patients with other diseases will help assess the models’ ability to differentiate hypertension from other health conditions.