Evidence level: B
“脉有常”与”脉应时辰”之争,本质是同一个问题:扣掉节律之后的脉搏频谱,到底比不扣更稳?这个看似技术性的问题,恰好是裁决两种中医时间观谁更接近物理现实的标尺。
正文
中医里有两句看似打架的话:一句说”脉有常”,一个人的脉象大体可辨,靠它能认人;另一句说”脉应时辰”,子午流注认为不同时辰脉象不同,《素问》明言”平旦人气生,日中阳气隆”。用频谱科学看,这两句其实在描述同一个对象的两个成分:脉搏频谱里既有跨时辰稳定的”身份层”,也有按时辰波动的”节律层”。真正的科学问题是:这两层能不能干净分开?把节律扣掉之后,”身份层”是不是更可辨?
先确认测量层是牢的。2015年王唯工学派的标准化谐波分析协议显示,22名健康人用同一套TD01C分析仪,操作者内重测的11个谐波指标ICC全部大于0.9(C1=0.97),操作者间ICC在0.83-0.96之间(PMID 27170904)。同年另一项研究用人工脉搏发生器检验仪器本身的内在信度,11个谐波成分的ICC全部大于0.95、CV全部小于3%,把仪器这一环也钉死了(PMID 25693606)。谐波谱的短期重测一致性在协议标准化后已无可争议——个体签名”稳不稳”的答案不在测量误差,而在真实生理节律的调制幅度。
再看节律层有多大。昼夜尺度上,1982年Halberg团队用24小时连续监测证明:血压、脉搏的日内变异系数达6.9%-17.9%,用余弦模型拟合,昼夜节律可解释SBP最多46%、脉搏最多42%的总变异(PMID 7176075)。2017年台湾团队用激光散斑血流图追踪视神经头脉搏波形,8个波形参数中6个有显著昼夜波动,阻力指数RI上午9点达峰、晚上9点落谷,节律可用三次多项式精确拟合(PMID 29095262)。季节尺度同样存在:意大利731例高血压患者中,颈股PWV与当日日照时长独立负相关(β=-0.170,p=0.001),冬天日照短、PWV高(PMID 28231259);台湾72名森林工作人员冬夏对照显示冬季baPWV、中心血压、体循环阻力全部显著升高(PMID 30429541)。节律不是噪声,是占频谱总变异40%以上的真实信号。
两个事实合起来,暴露了一个从未被检验的推论:既然节律解释了谱中40%以上的变异,而短期重测一致性又足够高,那么把节律参数化后扣除,剩下的”节律校正谱”应该比原始谱更稳定——个体签名不是被节律”污染”,而是被”包裹”,拆开反而更清晰。这正是源文章《脉搏频谱个体签名在时辰与季节尺度上的稳定性》”待验证”部分的第1项,至今没有直接研究。
可执行设计并不复杂:同一批健康人(60-100人)在昼夜×多季两个尺度上重复测量——每个季节窗口内选4-6个时辰各测一次,跨春、夏、秋、冬四轮,同一套标准化协议、同一时段窗避免日间混杂。分析分两步:第一步用谐波回归(单/多分量余弦,甚至三次多项式)对C1-C10每个谐波阶次拟合”谐波值-时辰”节律曲线,得到每人的节律参数(MESOR、幅度、峰相位);第二步用”实测值-节律预测值”构造节律校正谱,头对头比较原始谱与校正谱的重测一致性——组内相关系数ICC、最小可检测差异MDC、跨个体可分性(判别AUC),再做方差分解看校正谱是否把”时辰节律+季节漂移”两项贡献显著压缩。
三种结局都是知识,各自指向不同的中医解读。结局一:校正谱ICC显著高于原始谱(如0.85→0.95)——”脉有常”在物理层面获胜,个体签名的稳定性边界在”扣掉节律之后”更清晰,临床脉诊应记录时辰并做节律归一。结局二:两者无显著差异——节律的个体内影响远小于个体间差异,”脉有常”直接成立,子午流注的临床权重需下调。结局三:校正后ICC反而下降——节律本身也是身份的一部分(”这个人的脉就爱上午浮下午沉”),正确的做法不是校正而是把节律参数本身当签名特征,中医”脉象兼参时辰”的原始实践反而更接近正确模型。
三个坑要提前堵住。一是测量条件混杂:每次测量前统一静息10分钟、禁咖啡茶、同一室温(22-24℃),否则拟合出的”节律”是伪节律。二是周期覆盖不全:只测白天不测夜间,余弦模型会系统性失真,至少覆盖12小时以上相位空间,季节轮次跨满四季才能分离”季节漂移”与”个体漂移”。三是多重比较:C1-C10十个谐波阶次每个都做ICC比较,必须用Bonferroni或FDR校正,否则假阳性淹没真实信号。样本量按ICC 0.85 vs 0.95、α=0.05、β=0.20估算约60-100人,纵向流失率按15%-20%预留。
做完这套设计,得到的不仅是”校正流程”,更是一张个体签名稳定性的边界地图:低阶谐波(C1-C4)若在校正后跨时点稳定,则”身份层”假设获得直接证据,支撑”脉可辨人”;高阶谐波(C5-C10)若被节律主导,则”状态层”假设成立,支撑”脉应时辰”。这张地图同时回答了中医两个古老命题各自的适用域,也让”节律校正”从拍脑袋的预处理变成有数据裁决的标准步骤——这正呼应2026年最新综述指出的方向:谐波分析要真正落地,必须建立标准化采集与处理管线(PMID 42490820)。
证据与局限
已验证: 标准化协议下桡动脉谐波短期重测ICC>0.9、仪器内在信度CV<3%(PMID 27170904、25693606);昼夜节律可解释血压/脉搏总变异30%-46%(PMID 7176075);脉搏波形参数存在可多项式拟合的昼夜节律(PMID 29095262);PWV等动脉力学参数随日照时长/季节显著变化(PMID 28231259、30429541)。
部分验证: “低阶谐波稳定=身份层、高阶谐波易变=状态层”的分工假设有数据支持但无直接机制研究;”节律校正谱重测一致性应优于原始谱”是推论,尚无直接测量。
待验证: 同一人分时多点(昼夜×多季)重复测量后,节律校正谱与原始谱的ICC/MDC头对头比较——个体签名稳定性的确切边界,目前没有任何直接研究。
相关阅读
证据等级: B
参考文献: PMID 27170904, PMID 25693606, PMID 7176075, PMID 29095262, PMID 28231259, PMID 30429541, PMID 42490820
免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。
Rhythm-Corrected Spectrum vs Raw Spectrum: Head-to-Head Validation of Individual Signature Stability
Evidence level: B
The debate between “pulse has constancy” and “pulse responds to time” essentially boils down to a single question: Is the pulse spectrum more stable after removing the rhythm than before? This seemingly technical issue serves as the metric for determining which of the two Traditional Chinese Medicine (TCM) temporal frameworks aligns more closely with physical reality.
Main Text
Two seemingly contradictory statements exist in TCM: one asserts “pulse has constancy,” implying that pulse characteristics are generally identifiable for individual recognition; the other states “pulse responds to time,” noting that Zi Wu Liu Zhu holds that pulse characteristics vary by hour, with the Su Wen explicitly stating “human qi is generated at dawn, and yang qi peaks at noon.” From the perspective of spectral science, these two statements describe two components of the same object: the pulse spectrum contains both a “identity layer” that is stable across hours and a “rhythm layer” that fluctuates by hour. The genuine scientific question is whether these two layers can be cleanly separated; specifically, is the “identity layer” more discernible after removing the rhythm?
First, confirm that the measurement layer is robust. The 2015 standardized harmonic analysis protocol from Wang Wei’s engineering school showed that among 22 healthy individuals using the same TD01C analyzer, the inter-classifier correlation coefficient (ICC) for all 11 harmonic indicators exceeded 0.9 (C1=0.97), and inter-operator ICCs ranged from 0.83 to 0.96 (PMID 27170904). In another study that year, an artificial pulse generator was used to test the instrument’s intrinsic reliability, showing ICCs greater than 0.95 and coefficients of variation (CV) less than 3% for all 11 harmonic components, effectively securing this link (PMID 25693606). Short-term retest consistency of the harmonic spectrum is undisputed following protocol standardization—the answer regarding whether the individual signature is “stable” does not lie in measurement error but in the modulation amplitude of real physiological rhythms.
Next, examine the magnitude of the rhythm layer. On a diurnal scale, Halberg’s team in 1982 demonstrated using 24-hour continuous monitoring that the intra-day coefficient of variation for blood pressure and pulse ranged from 6.9% to 17.9%. Fitting a cosine model revealed that circadian rhythms explained up to 46% of total variation in systolic blood pressure (SBP) and up to 42% in pulse (PMID 7176075). In 2017, a Taiwanese team tracked pulse waveforms at the optic nerve head using laser speckle flowgraphy; among 8 waveform parameters, 6 showed significant diurnal fluctuations, with resistance index (RI) peaking at 9 AM and troughing at 9 PM, a rhythm precisely fitted by a cubic polynomial (PMID 29095262). Seasonal scales also exhibit variation: among 731 hypertensive patients in Italy, carotid-femoral pulse wave velocity (PWV) showed an independent negative correlation with daily daylight duration ($\beta=-0.170$, $p=0.001$), with higher PWV in winter due to shorter daylight (PMID 28231259); a Taiwanese comparison of 72 forest workers between winter and summer showed significant increases in winter baPWV, central blood pressure, and total peripheral resistance (PMID 30429541). Rhythm is not noise; it is a real signal accounting for over 40% of total spectral variation.
These two facts together expose an inference never directly tested: since rhythm explains over 40% of variation in the spectrum, yet short-term retest consistency remains high, parameterizing and subtracting the rhythm should yield a “rhythm-corrected spectrum” that is more stable than the raw spectrum—the individual signature is not “polluted” by rhythm but “enveloped,” and separating it yields greater clarity. This corresponds to Item 1 of the “pending verification” section in the source article Stability of Pulse Spectrum Individual Signatures at Circadian and Seasonal Scales, which remains unstudied to date.
The executable design is not complex: repeat measurements on the same cohort of healthy individuals (60–100 people) across two scales (day-night × multiple seasons)—selecting 4–6 hours within each seasonal window for four rounds across spring, summer, autumn, and winter, using the same standardized protocol and time windows to avoid diurnal confounding. Analysis proceeds in two steps: First, fit “harmonic value vs. hour” rhythm curves for each harmonic order (C1–C10) using harmonic regression (single/multi-component cosine, or cubic polynomial), deriving individual rhythm parameters (MESOR, amplitude, peak phase); Second, construct the rhythm-corrected spectrum using “measured value minus rhythm predicted value” and perform a head-to-head comparison of retest consistency between the raw spectrum and corrected spectrum—comparing ICCs, minimum detectable differences (MDC), and cross-individual separability (discrimination AUC), followed by variance decomposition to assess whether the corrected spectrum significantly compresses contributions from “diurnal rhythm” and “seasonal drift.”
All three outcomes constitute knowledge, pointing toward different TCM interpretations. Outcome 1: Corrected spectrum ICC significantly exceeds raw spectrum ICC (e.g., 0.85 → 0.95)—”pulse has constancy” wins at the physical level; the stability boundary of individual signatures is clearer after removing rhythm, suggesting clinical pulse diagnosis should record time and apply rhythm normalization. Outcome 2: No significant difference between the two—individual-intra-rhythm effects are far smaller than individual-differences, supporting “pulse has constancy” directly and warranting a downward adjustment of the clinical weight of Zi Wu Liu Zhu. Outcome 3: Corrected ICC decreases—rhythm itself is part of identity (“this person’s pulse prefers floating in the morning and sinking in the afternoon”); the correct approach is not correction but treating rhythm parameters as signature features, suggesting the original TCM practice of “pulse diagnosis combined with time reference” aligns more closely with the correct model.
Three pitfalls must be preemptively addressed. First, measurement condition confounding: standardize 10 minutes of rest, prohibit coffee/tea, and maintain room temperature (22–24°C) before each measurement; otherwise, the fitted “rhythm” will be spurious. Second, incomplete cycle coverage: measuring only daytime without nighttime will cause systematic distortion in cosine models; at minimum cover a 12-hour phase space, and span all four seasons to separate “seasonal drift” from “individual drift.” Third, multiple comparisons: performing ICC comparisons for all ten harmonic orders (C1–C10) requires Bonferroni or FDR correction; otherwise, false positives will drown out true signals. Sample size estimation for ICC 0.85 vs. 0.95, $\alpha=0.05$, $\beta=0.20$ suggests 60–100 individuals, with longitudinal dropout reserved at 15%–20%.
Completing this design yields not only a “correction process” but also a boundary map of individual signature stability: if lower-order harmonics (C1–C4) remain stable across time points after correction, the “identity layer” hypothesis gains direct evidence supporting “pulse can identify individuals”; if higher-order harmonics (C5–C10) are dominated by rhythm, the “state layer” hypothesis holds, supporting “pulse responds to time.” This map simultaneously addresses the respective applicability domains of TCM’s two ancient propositions and transforms “rhythm correction” from a heuristic preprocessing step into a data-decided standard procedure—echoing the direction pointed out by the latest 2026 review: for harmonic analysis to truly land, standardized acquisition and processing pipelines must be established (PMID 42490820).
Evidence and Limitations
Verified: Under standardized protocols, radial artery harmonic short-term retest ICC > 0.9 and instrument intrinsic reliability CV < 3% (PMID 27170904, 25693606); diurnal rhythms explain 30%–46% of total variation in blood pressure/pulse (PMID 7176075); pulse waveform parameters exhibit diurnal rhythms fit by polynomials (PMID 29095262); arterial mechanics parameters like PWV vary significantly with daylight duration/season (PMID 28231259, 30429541).
Partially Verified: The division hypothesis (“lower-order harmonics stable = identity layer, higher-order harmonics variable = state layer”) has data support but lacks direct mechanistic research; “rhythm-corrected spectrum retest consistency should exceed raw spectrum” is an inference with no direct measurement yet.
Pending Verification: Head-to-head comparison of ICC/MDC between rhythm-corrected and raw spectra after repeated multi-point measurements (day-night × multiple seasons) on the same individual—the exact boundary of individual signature stability currently has no direct research.
Related Reading
- Stability of Pulse Spectrum Individual Signatures at Circadian and Seasonal Scales
- Longitudinal Stability and Plasticity of Pulse Spectrum Individual Signatures
- Zi Wu Liu Zhu = Frequency Encoding of Circadian Rhythms
- Quantitative Mapping Verification of Harmonic Correspondence to Organs
- Research Pool: Bio-frequency Codes
Evidence Level: B
References: PMID 27170904, PMID 25693606, PMID 7176075, PMID 29095262, PMID 28231259, PMID 30429541, PMID 42490820
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.
Leave a Reply