Author: joe

  • 同一个「稳」字,两套算法给出两个相反的排名:φ 的圆形离散度与算术 SD 在跨零呼吸上的偏差标定

    Evidence level: C

    两套算法,两个排名。吃亏的偏偏是练得最稳的那一组——因为偏差的大小,本身就是由「谁更稳」决定的。

    一、先把切点说清:这篇拆的不是 MDC,是排名

    531 那一节把「修炼者呼吸更稳」挂在 φ 的离散度上。此前已有两篇动过这把尺子:circular-mdc-cross-zero-arithmetic-artifact-calibration 拆的是单个差值的 MDC 公式在跨零时失效,判的是「测得出测不出」;dantian-phase-dispersion-stability-vs-freezing 拆的是离散度大到底算好还是算坏,判的是「解读方向」。这一篇问第三件事:两套算法给出的数,会不会把两组的排名对调。这是三个不同的问题。前两个都判好了,排名照样可能是反的——排名只看次序,而偏差恰好与「谁更同步」共线。

    二、偏差不是随机的,它和 φ 均值同向

    算术 SD 把 −179° 与 +179° 当成相隔 358°,实际只隔 2°。关键在于偏差方向与大小取决于 μ 落在哪:μ 越靠近 0°(几乎同相),越容易有几次相位从正翻负,样本 SD 就被这几次翻转撑大。于是偏差在「同步最好的人」身上最大。这不是噪声,是系统性的、与自变量共线的偏移:你要比较的正是「谁更同步」,而偏差强度恰好由同步程度决定。方向也对不上——偏差让人显得更不稳,实际更稳。圆形侧同样不干净:1−R(即相位锁定值、Kuramoto 序参量)在完全随机时只有样本量趋于无穷才趋于零,有限样本下它的期望值随 n 变化,跨样本量比较必然带偏(PMID 37756609)。该文的重归一化版本正为此设计:让期望值不再依赖 n。

    三、更狠的一层:样本量恰好和「呼吸慢」锁死

    第二节的两条偏差本来可以分开处理,偏偏在这里被同一个设计绑死。静息记录通常固定时长(如 10 分钟),呼吸次数 n 与呼吸周期成反比。练得深的人呼吸慢,10 分钟里凑不满次数。而 1−R 的负偏在 n 小时更大——于是「呼吸慢」这个本该是功力的标志,通过 n 变成了「离散度偏大」的伪信号。算术 SD 侧同理:n 小则单次跨零翻转的权重更大。这就构成了双重共线:偏差既随 μ 变、又随 n 变,而 μ 与 n 在修炼者身上同向变动。两条偏差叠起来,方向一致地指向「练得越深越显得不稳」。这比单纯说「算法有偏」严重:既往任何用固定时长窗口、未报 φ 均值与呼吸次数的组间比较,结论方向都不可预判。

    四、RIP 这层还得再剥一次

    上面都在讨论「拿到 φ 序列之后怎么算」,可 φ 本身怎么来的也有讲究。以 Lissajous 图算相位角是 TAA 的常规做法,但该法有数学限制:2290 个周期的分析显示,相位角只在 −86.90° 到 +88.40° 之间取值,而全局相位延迟覆盖 −179.75° 到 +178.54°;相位角与环面积无法正确分类严重偏移与矛盾运动(PMID 38829281)。即当不同步严重到出现矛盾呼吸时,算术相位角自己饱和,把「很不同步」压成上限值。对做组间比较的人,这是方向明确的坑:组间差被压扁在最不同步那一端。再叠一个实测参照:儿童 PSG 中 θavg 在 N3 期为 13.7°、REM 期为 33.1°,差 2.4 倍(PMID 29991414)——同一个人换个睡眠分期就能翻倍,跨组比较若不同分期,根本谈不上可比。

    五、还有一层:误差本身可以「分组不同」

    即便离散度算对了,组间差仍可能偏。差异测量误差(differential measurement error)指误差的性质随组别或时点而变,模拟研究显示它会让处理效应估计产生偏倚、并降低检出功效、影响置信区间覆盖率(PMID 34530859)。放到这里:新手戴着带子紧张、动作多、基线漂移大;老手躺下不动。误差结构天然分组不同,而 φ 的噪声底本来就不小——早产儿安静睡眠期 φ 的个体内 CV 达 11%–123%,且经机械模型排除后判定源于相位关系本身不稳,而非算法(PMID 24716708)。分母这么大,再叠一层分组误差,组间差里有多少是「稳」、多少是「谁更会配合测量」,分不开。

    六、判据与做法

    做法分两段。合成段:Monte Carlo 造 φ 序列,真实集中度 κ 已知,μ 从 0° 扫到 ±60°,n 取 20/50/200(对应不同呼吸速率),含与不含跨零两情形,比较算术 SD 与圆形离散度(1−R、圆方差、von Mises κ)对真实 κ 的恢复偏差。真实段:≥90 人同体位 ≥10 分钟 RIP,双算法重算,看排名是否翻转,并报 φ 均值与呼吸次数。判据三格:① 合成中算术 SD 在 μ→0 时相对真实离散度高估 >30%、随 μ→0 单调放大,且 n=20 与 n=200 之间差异 >10 个百分点 ⇒ 算法伪迹与样本量偏差双重确认,未报 φ 均值与 n 的 SD 结论一律不可解释;② 真实数据上双算法排名出现翻转(哪怕只一对组)⇒ 既往以算术 SD 论证「修炼者更稳/更不稳」的文献全部须重算;③ 两算法排名一致且偏差 <10% ⇒ 算术 SD 在成人静息 φ 上可继续使用,本文的方法学担忧撤回。必须自认:30% 与 10 个百分点均为本轮自定、无文献出处,执行前须改成以 MDC95 为锚的阈值。

    七、这一篇能带来什么

    价值不在「又发现一个算法缺陷」,而在把「稳」的可比性问题推到排名层面。离散度是相对量,排名是它唯一用法;偏差与自变量共线,「更稳」这个比较本身就在比偏差。更朴素的教训是:圆形数据的无偏度量早就存在(成对相位一致性即为一例,PMID 20114076),但呼吸协调领域大量沿用的仍是算术 SD。不是不知道,是没人在比较前算过这笔账。对丹田研究的直接含义是:在把 φ-SD 当终点之前,先做一次纯算法的偏差标定,比再招一批受试者便宜。

    证据与局限

    已验证:圆形量的集中度指标(1−R/相位锁定值/序参量)在有限样本下期望值随 n 变化、跨样本量比较带偏,已有重归一化方案(PMID 37756609);Lissajous 相位角存在取值饱和与分类失效(PMID 38829281);θavg 随睡眠分期显著变化(PMID 29991414);差异测量误差可致效应估计偏倚与功效下降(PMID 34530859)。

    部分验证:φ 的噪声底量级有实测支撑(CV 11%–123%,PMID 24716708),但该数据来自早产儿,成人静息是否同量级未知;圆形统计的无偏替代度量存在且有工程可用性(PMID 20114076),但在呼吸协调领域尚未被系统采用。

    待验证:算术 SD 在 μ→0 时的高估量级、以及该高估与 n 的交互大小,全部是本文推论,尚无任何实测标定;真实 RIP 数据上双算法排名是否真的会翻转,完全没有数据;「呼吸慢 ⇒ n 小 ⇒ 显得不稳」这条三重共线的实际效应量,也未测。

    相关阅读

    证据等级: C
    参考文献: PMID 37756609, PMID 38829281, PMID 29991414, PMID 34530859, PMID 24716708, PMID 20114076
    L1 网络状态:CORRESPONDENCE_NETWORK.md 2.2 物理侧仅列「相位 / 节律 | — | 周期性耦合的关键」,2.3 生理侧仅列「呼吸率 / 呼吸模式 | 可穿戴 | 与节律耦合最直接」;逐词检索「离散度/圆形统计/算术/排名/样本量/偏差/方差/算法」全部 0 命中,边表 3.1–3.5 止于 E-050(层级 L1、状态 open,现代证据栏原文「尚无任何可测来源」),无任何以「离散度 / 算法 / 排名」为源或目标的边 —— 该主题尚未进入 L1 对应网络。


    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。

    The same \”steady\” character, two algorithms yielding opposite rankings: deviation calibration of φ’s circular dispersion versus arithmetic SD across zero-breathing

    Evidence level: C

    Two algorithms, two rankings. The group that trained the most stably loses out—because the magnitude of the bias itself is determined by “who is more stable.”

    I. Clarify the Cut-Point First: This Paper Does Not Split MDC, It Splits the Ranking

    Section #531 hangs “the practitioner’s breathing is more stable” on φ’s dispersion. Two prior papers have already wielded this ruler: circular-mdc-cross-zero-arithmetic-artifact-calibration dissects the MDC formula for a single difference when crossing zero, judging “measurable vs. unmeasurable”; dantian-phase-dispersion-stability-vs-freezing dissects whether large dispersion is good or bad, judging the “interpretation direction.” This paper asks the third question: Will the numbers from the two algorithms flip the ranking of the two groups. These are three distinct questions. The first two have been judged, yet the ranking can still be reversed—ranking depends solely on order, while the bias is collinear with “who is more synchronized.”

    II. The Bias Is Not Random; It Aligns with the φ Mean

    Arithmetic SD treats −179° and +179° as separated by 358°, whereas they are actually only 2° apart. The key lies in the direction and magnitude of the bias depend on where μ falls: the closer μ is to 0° (near perfect synchrony), the more likely several phases flip from positive to negative, inflating the sample SD due to these flips. Thus, the bias is largest for “the most synchronized person.” This is not noise; it is a systematic, collinear shift with the independent variable: you are comparing “who is more synchronized,” yet the bias intensity is exactly determined by the degree of synchronization. The direction also mismatches—the bias makes people appear less stable, while they are actually more stable. The circular side is similarly flawed: 1−R (i.e., phase locking value, Kuramoto order parameter) only approaches zero as sample size tends to infinity under complete randomness; under finite samples, its expected value varies with n, so cross-sample-size comparisons inevitably introduce bias (PMID 37756609). The paper’s re-normalized version was designed precisely for this: to make the expected value independent of n.

    III. An Even Sharper Layer: Sample Size Is Locked to “Slow Breathing”

    The two biases in Section II could have been handled separately, yet here they are bound together by the same design. Resting recordings typically have fixed duration (e.g., 10 minutes); the number of breaths n is inversely proportional to the breathing cycle. Those who train deeper breathe slower, so within 10 minutes they cannot complete enough breaths. Meanwhile, the negative bias of 1−R grows larger as n decreases—thus “slow breathing,” which should be a marker of mastery, becomes via n a false signal of “large dispersion.” The arithmetic SD side is similar: smaller n gives greater weight to single zero-crossings. This creates double collinearity: the bias depends both on μ and on n, yet μ and n move in the same direction for practitioners. Stacked together, these two biases consistently point toward “the deeper the training, the less stable they appear.” This is worse than simply stating “the algorithm is biased”: any prior between-group comparison using fixed-duration windows without reporting φ mean and breathing count has an unpredictable conclusion direction.

    IV. RIP Requires One More Peeling Layer

    The above discusses “how to compute after obtaining the φ sequence,” but how φ itself is derived also matters. Computing phase angles via Lissajous plots is standard in TAA, yet this method has mathematical limits: analysis of 2,290 cycles shows phase angles range only from −86.90° to +88.40°, while global phase delay spans −179.75° to +178.54°; phase angle and ring area cannot correctly classify severe offsets and contradictory motion (PMID 38829281). That is, when asynchrony becomes so severe that contradictory breathing occurs, the arithmetic phase angle saturates, compressing “very asynchronous” into its upper limit. For those doing between-group comparisons, this is a pitfall with clear direction: between-group differences are compressed at the most asynchronous end. Adding an empirical reference: in pediatric PSG, θavg is 13.7° in N3 stage and 33.1° in REM stage, a difference of 2.4× (PMID 29991414)—the same person can double across sleep stages, making cross-group comparison incomparable if stages differ.

    V. Another Layer: Error Structure Itself Can Differ by Group

    Even if dispersion is computed correctly, between-group differences may still be biased. Differential measurement error refers to error properties varying by group or time point; simulation studies show it biases effect estimates, reduces detection power, and affects confidence interval coverage (PMID 34530859). Applied here: novices wearing straps feel tense, move more, and have larger baseline drift; experts lie still. The error structure is inherently group-specific, while φ’s noise floor is already substantial—premature infants’ quiet sleep phase φ has an individual CV of 11%–123%, and after mechanical model exclusion, this is judged to stem from inherent phase instability rather than algorithmic issues (PMID 24716708). With such a large denominator, stacking another layer of group-specific error makes it impossible to distinguish how much of the between-group difference reflects “stability” versus “who cooperates better with measurement.”

    VI. Criteria and Methods

    The approach has two segments. Synthetic segment: Monte Carlo generates φ sequences with known true concentration κ, μ sweeps from 0° to ±60°, n takes values of 20/50/200 (corresponding to different breathing rates), including both with and without zero-crossing, comparing arithmetic SD against circular dispersion (1−R, circular variance, von Mises κ) in terms of recovery bias from true κ. Real segment: ≥90 individuals in same posture ≥10 minutes of RIP, recompute with both algorithms to check if rankings flip, and report φ mean and breathing count. Three criteria: ① In synthetic data, arithmetic SD overestimates true dispersion by >30% as μ→0, monotonically increasing, and the difference between n=20 and n=200 exceeds 10 percentage points ⇒ confirms both algorithmic artifact and sample-size bias; any SD conclusion not reporting φ mean and n is uninterpretable; ② In real data, if both algorithms flip rankings (even just one pair of groups) ⇒ all prior literature using arithmetic SD to argue “practitioners are more stable/less stable” must be recalculated; ③ If both algorithms agree and bias <10% ⇒ arithmetic SD can continue for adult resting φ, and this methodological concern is withdrawn. Must self-admit: the 30% and 10 percentage points are self-defined thresholds for this round, with no literature source; before execution, they must be revised to use MDC95 as the anchor.

    VII. What This Paper Can Bring

    The value lies not in “discovering another algorithmic defect,” but in pushing the comparability of “stability” to the ranking level. Dispersion is a relative quantity, and ranking is its sole usage; bias is collinear with the independent variable, so “more stable” itself compares against the bias. A simpler lesson is: unbiased measures for circular data already exist (paired phase consistency is one example, PMID 20114076), yet the breathing coordination field largely continues using arithmetic SD. They are not unaware; no one has calculated this account before comparison. For dantian research, the direct implication is: before taking φ-SD as the endpoint, perform a pure algorithmic bias calibration—cheaper than recruiting another batch of subjects.

    Evidence and Limitations

    Verified: Concentration metrics for circular quantities (1−R/phase locking value/order parameter) have expected values varying with n under finite samples, causing cross-sample-size comparison bias, and a re-normalization solution exists (PMID 37756609); Lissajous phase angles exhibit saturation and classification failure (PMID 38829281); θavg varies significantly across sleep stages (PMID 29991414); differential measurement error can bias effect estimates and reduce power (PMID 34530859).

    Partially Verified: φ’s noise floor magnitude has empirical support (CV 11%–123%, PMID 24716708), but this data comes from premature infants, leaving unknown whether adults at rest share the same scale; unbiased alternatives for circular statistics exist and are engineering-ready (PMID 20114076), yet not systematically adopted in the breathing coordination field.

    To Be Verified: The degree of overestimation by arithmetic SD as μ→0, and the magnitude of interaction between this overestimation and n, are entirely inferred in this paper with no empirical calibration; whether rankings truly flip under both algorithms on real RIP data lacks any data; the actual effect size of “slow breathing ⇒ small n ⇒ appears unstable” remains unmeasured.

    Related Reading

    Evidence Level: C
    References: PMID 37756609, PMID 38829281, PMID 29991414, PMID 34530859, PMID 24716708, PMID 20114076
    L1 Network Status: CORRESPONDENCE_NETWORK.md Section 2.2 Physical side lists only “phase/rhythm | — | key periodic coupling”; Section 2.3 Physiological side lists only “breathing rate/breathing pattern | wearable | most directly coupled to rhythm”; keyword search for “dispersion/circular statistics/arithmetic/ranking/sample size/bias/variance/algorithm” yields zero hits; edge table 3.1–3.5 stops at E-050 (layer L1, status open, original text states “no measurable source yet“), with no edges originating from or targeting “dispersion/algorithm/ranking” — this topic has not yet entered the L1 corresponding network.


    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.

  • 场次噪声就能伪造出十几年功力:年限斜率的两个方差成分,与横断面替代纵向的功效陷阱

    Evidence level: C

    「练得越久 Δ 越窄」这句话里藏着两次减法。前一次是 Δ = 练功态 − 静息态,后一次是 Δ 对年限回归。把第一次减法做对,救不了第二次。

    一、两个减法,前一篇只拆了第一个

    两态差值 Δ 的测量误差,此前已经单独审过一次:两个成分各自一致,不代表相减之后还一致,MDC95 那条闸门就是为此设的。

    但那只是第一道。第二道减法从来没被拆开过:把 Δ 对修炼年限做回归,看斜率是否显著为负。这一步引入两个全新的误差源——场次之间的随机抖动,以及年限分组本身的混杂。前者与”测得准不准”无关(单个 Δ 可以很准而组间差全是噪声),后者与样本量无关(招再多人也消不掉)。一个设计可以在第一道减法上完全合格,却在第二道减法上从一开始就测错了对象。

    二、场次噪声恰好和几十年年龄差一个量级

    这是整件事最要命的一个数字。一项针对成人知觉—运动表现的重复测量研究给出:场次之间的变异,平均相当于个体间变异的 25%–50%;更关键的是后半句——这个量级与横断面比较中跨越 12 至 27 岁年龄范围所显示的差异相当(PMID 15014087)。

    把它换成本题的语言:如果你按”新手 / 1–5 年 / 5–10 年 / 10 年以上”分层,那么同一个人隔天再测一次的抖动,就足以伪造出一个相当于”多练了十几年”的组间差。

    这不是”噪声有点大需要多招人”,而是效应量与噪声底重叠。在这种情况下,任何按年限分层的横断面结果,都无法与”场次噪声恰好这么排”相区分。

    三、斜率的方差有两个来源,不是一个

    方法学上的正确骨架早在三十多年前就写清楚了:个体斜率(对时间回归,取最小二乘估计)的方差包含两个成分——一个来自个体内变异(由测量误差、随访长度与测量频率共同决定),另一个是真实的个体间斜率差异(PMID 2362980)。

    这条公式对本题是致命的。随访期越短、次数越少,个体内那一项就越大。而一个”四层年限、每人只测一次”的横断面设计,本质上把个体内成分推到了最大,真实斜率差异那一项被埋在下面。

    所以功效计算不能拿”总方差”当分母——必须把两项分开报。只报总方差的设计,等于承认自己不知道算的是哪一项。

    四、用横断面替代纵向,是这道题最贵的捷径

    为什么大家爱用横断面?因为便宜,但便宜有代价,代价可以量化。同一批影像遗传学数据上,纵向分析与仅用基线的横断面分析相比,在全基因组显著性水平上检出的关联更多(PMID 25098835)。反向推论就是:横断面要达到同等功效,需要的样本更多。而按变化率比较的样本量公式还多要一个参数——暴露的组内相关;时变暴露的研究在比较变化率时普遍不如时不变暴露高效(PMID 19899065)。本题的”修炼年限”恰恰是最典型的时变暴露。所以四层横断面设计的功效问题不是”够不够”,而是分子从一开始就错了。

    五、短期和长期斜率,可以指向相反方向

    即便改成纵向,还有一层:短期变化与长期变化不是同一个东西。

    三层多层模型把二者同时参数化后,出现了拆分才看得见的分裂——简单选择反应时的反应时不一致度短期平稳、跨年上升;而 one-back 任务的不一致度短期显著下降、长期无变化(PMID 35721020)。同一个指标,两个时间尺度的斜率方向相反,合成一个斜率就等于互相抵消。

    对应到本题:Δ 随年限的”变化”里,如果混着”第一次测的新鲜感”和”练了十年之后的结构改变”,这两股力的方向可能不同。不先分开参数化,得到的合成斜率没有解释力。

    六、训练效应会把趋势读反,不只是读小

    更坏的情形是符号错。

    一项 4718 人的纵向研究里,认知得分在第一年逐步改善,此后才开始恶化;调整训练效应后,组间差异才达到显著(p = 0.019),而”整体衰退”的表观趋势消失(PMID 11901278)。也就是说,不调整时看到的趋势和调整后的趋势,可以完全不是一回事。

    而训练效应与真实变化在常规设计里几乎不可分:等间隔随访中,时间与暴露次数完美共线,二者只有在特定收敛假设下才能分离;假设不成立时,偏倚与一类错误率同时失控(PMID 21639642)。

    对本题,这条曲线的形状可以预期:新手头几次 Δ 快速上升(学会了怎么鼓肚子),老手早已平台。这条曲线若被读成年限效应,符号正好是反的。

    七、年限与年代,最后一刀切不开

    最底层的问题不在测量,在识别。

    年龄、时期、世代三者满足”世代 = 时期 − 年龄”的线性依赖,线性成分因此无法被分别估计;各种约束方案仍在争论,模拟显示没有任何一种模型能通吃(PMID 40779571)。随机效应方案也不占便宜——它施加的约束是任意的、且高度隐晦,与其他受约束估计量并无本质差别(PMID 36644415)。

    落到本题:一个六十岁、练了三十年的人,他的 Δ 窄,可能因为练得久,也可能因为他出生在那一代人里、那批人本来就不同。四层年限设计对此完全无解。

    证据与局限

    已验证: 场次间变异可达个体间变异的 25%–50% 且等量于跨 12–27 岁的横断面差(PMID 15014087);斜率方差由个体内与个体间两成分构成(PMID 2362980);纵向分析在同批数据上功效优于横断面(PMID 25098835);短期与长期斜率可反向(PMID 35721020);训练效应可反转趋势读法(PMID 11901278);APC 线性成分不可分离(PMID 40779571;PMID 36644415)。这些均为方法学层面的定论级证据。

    部分验证: 上述数字全部来自认知与知觉—运动领域,没有一项是在胸腹相位角或腹区占比上做的。跨领域外推的方向可信(方差成分结构是普适的),但量级不能直接搬——φ 的个体内变异系数在不同人群间跨 11%–123%,本身离散极大。

    待验证: 三个关键量在本主题上完全空白——① Δ 的个体内与个体间方差各占多少(决定年限斜率是否存在可测空间);② φ/%AB 上的 retest gain 有多大、几次之后饱和;③ 修炼年限与出生年代在当前修炼人群中的共线强度。这三项不测出来,本文的判据只能算”条件性成立”。

    L1 网络状态:该主题尚未进入 L1 对应网络。CORRESPONDENCE_NETWORK.md 节点表按「易学/中医侧」「物理侧」「生理/体验侧」三类域分派,逐词检索「信度 / 重测 / 最小可检测 / MDC / 功效 / 效度 / 胸腹 / 年限 / 年资 / ICC / 样本量」全部 0 命中;边表止于 E-050(「六十甲子纳音 | 对应 | 60 个频率」,层级 L1、状态 open,现代证据栏原文「尚无任何可测来源」),无任何以方差成分、统计功效或年限斜率为源或目标的边。EDGE_TABLE_science-yixue.md 仅有概念计数(丹田 51 篇 / 气功 48 篇 / 呼吸 189 篇),无 L1/L2/L3 层级与 active/open/conflict 判定。

    相关阅读

    证据等级: C
    参考文献: PMID 15014087, PMID 2362980, PMID 25098835, PMID 19899065, PMID 35721020, PMID 11901278, PMID 21639642, PMID 40779571, PMID 36644415


    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。

    Session Noise Can Forge Decades of Proficiency: The Two Variance Components of Slope Rates and the Cross-Sectional Longitudinal Efficacy Trap

    Evidence level: C

    The phrase “the longer one practices, the narrower Δ becomes” conceals two subtractions. The first is Δ = practice state − resting state; the second is the regression of Δ against tenure. Getting the first subtraction right does not save the second.

    I. Two Subtractions; the Previous Article Only Decomposed the First

    The measurement error of the two-state difference Δ was already scrutinized separately: each component is consistent, yet their difference need not be, and the MDC95 threshold was set precisely for this reason.

    That is only the first cut. The second subtraction has never been decomposed: regressing Δ against years of practice to test whether the slope is significantly negative. This step introduces two entirely new error sources—random jitter between sessions and confounding inherent in tenure grouping. The former is unrelated to “how accurately we measure” (a single Δ can be precise while group differences are pure noise); the latter is unrelated to sample size (no matter how many recruits, confounding cannot be eliminated). A design can be fully adequate for the first subtraction yet measure the wrong object from the start for the second.

    II. Session Noise Matches the Magnitude of Decades of Age Difference

    This is the most critical number in the whole matter. A repeated-measures study of adult perceptual–motor performance reports: between-session variability averages 25%–50% of between-individual variability; more crucially, this magnitude matches the differences observed across a 12–27 year age range in cross-sectional comparisons (PMID 15014087).

    Translating to this topic’s language: if you stratify by “novice / 1–5 years / 5–10 years / 10+ years,” then the jitter from retesting the same person the next day is sufficient to fabricate a group difference equivalent to “practicing for an extra decade or more.”

    This is not “noise is too large; we need more recruits.” It is effect size overlapping the noise floor. Under these conditions, any cross-sectional result stratified by tenure cannot be distinguished from “session noise happening to align exactly this way.”

    III. The Slope Variance Has Two Sources, Not One

    The methodological skeleton was laid out over thirty years ago: the variance of individual slopes (regressed on time, estimated via least squares) comprises two components—one from within-individual variation (determined by measurement error, follow-up duration, and measurement frequency), and the other from true between-individual slope differences (PMID 2362980).

    This formula is fatal for this topic. Shorter follow-up and fewer measurements inflate the within-individual term, burying the true between-individual slope differences.

    Therefore, power calculations cannot use “total variance” as the denominator—the two terms must be reported separately. Designs reporting only total variance admit ignorance of which component they are actually estimating.

    IV. Replacing Longitudinal with Cross-Sectional Is the Most Expensive Shortcut

    Why do people favor cross-sectional designs? Because they are cheap, but cheapness has a quantifiable cost. On the same imaging-genetics dataset, longitudinal analysis detects more associations at genome-wide significance than cross-sectional analysis using baseline only (PMID 25098835). In reverse inference: cross-sectional designs require larger samples to achieve equivalent power. Moreover, sample size formulas based on rate-of-change comparisons demand an additional parameter—the intra-group correlation of the exposure; studies with time-varying exposures are generally less efficient than those with time-invariant exposures (PMID 19899065). “Years of practice” is precisely the quintessential time-varying exposure. Thus, the power problem of a four-tier cross-sectional design is not “is it enough?” but the numerator is wrong from the start.

    V. Short-Term and Long-Term Slopes Can Point in Opposite Directions

    Even if switched to longitudinal, another layer remains: short-term change and long-term change are not the same thing.

    Three-level multilevel modeling, which parameterizes both simultaneously, reveals a split visible only upon decomposition: simple choice reaction time shows short-term stability with year-over-year increases; whereas the one-back task shows short-term significant decline with no long-term change (PMID 35721020). For the same metric, slopes over two time scales point in opposite directions; combining them into a single slope results in mutual cancellation.

    Mapped to this topic: within the “change” of Δ over tenure, if it mixes “freshness from the first measurement” and “structural change after ten years of practice,” these forces may oppose each other. Without separate parameterization, the composite slope lacks interpretability.

    VI. Training Effects Can Reverse the Trend, Not Just Diminish It

    An even worse scenario involves sign reversal.

    In a longitudinal study of 4,718 participants, cognitive scores improved gradually in the first year before beginning to decline; after adjusting for training effects, between-group differences reached significance (p = 0.019), while the apparent trend of “overall decline” vanished (PMID 11901278). That is, the unadjusted trend and the adjusted trend can be entirely different.

    In conventional designs, training effects and true change are nearly inseparable: in equidistant follow-ups, time and exposure frequency are perfectly collinear, separable only under specific convergence assumptions; when those fail, both bias and Type I error rates run wild (PMID 21639642).

    For this topic, the expected curve shape is: Δ rises rapidly for novices (learning how to expand the abdomen), while experts are already plateaued. If this curve is read as a tenure effect, the sign is exactly reversed.

    VII. Tenure and Calendar Year: The Final Cut Cannot Be Made

    The deepest issue lies not in measurement but in identification.

    Age, period, and cohort satisfy the linear dependency “cohort = period − age”; thus linear components cannot be separately estimated; various constrained approaches remain debated, with simulations showing no model universally applicable (PMID 40779571). Random effects approaches offer no advantage—their constraints are arbitrary and highly implicit, differing in essence from other constrained estimators (PMID 36644415).

    Back to this topic: a 60-year-old who has practiced for thirty years may show narrow Δ either because of long practice or because they were born into a cohort with inherently different characteristics. A four-tier tenure design is completely unsolvable for this.

    Evidence and Limitations

    Verified: Between-session variability can reach 25%–50% of between-individual variability and equals cross-sectional differences across ages 12–27 (PMID 15014087); slope variance comprises within-individual and between-individual components (PMID 2362980); longitudinal analysis outperforms cross-sectional on the same dataset (PMID 25098835); short-term and long-term slopes can oppose each other (PMID 35721020); training effects can reverse trend interpretation (PMID 11901278); APC linear components are inseparable (PMID 40779571; PMID 36644415). These are methodological conclusions.

    Partially Verified: All cited numbers come from cognitive and perceptual–motor domains; none address thoraco-abdominal phase angle or abdominal proportion. Cross-domain extrapolation is credible (variance component structure is universal), but magnitude cannot be directly transferred—φ’s within-individual coefficient of variation spans 11%–123% across populations, showing extreme heterogeneity.

    To Be Verified: Three key quantities are entirely blank for this topic: ① the split between within-individual and between-individual variance of Δ (determining whether a measurable tenure slope exists); ② the retest gain on φ/%AB and after how many repetitions saturation occurs; ③ the collinearity strength between years of practice and birth cohort among current practitioners. Without these, this article’s criteria can only be deemed “conditionally valid.”

    L1 Network Status: This topic has not yet entered the L1 correspondence network. The CORRESPONDENCE_NETWORK.md node table, distributed across “Yijing Studies/TCM side,” “physics side,” and “physiological/experiential side,” yields zero hits for all terms: “reliability / retest / minimum detectable / MDC / power / validity / thoraco-abdominal / tenure / seniority / ICC / sample size.” The edge table stops at E-050 (“Sixty Jiazi Na Yin | Correspondence | 60 Frequencies,” level L1, status open; modern evidence column reads “no measurable source yet“), with no edges sourced or targeted by variance components, statistical power, or tenure slope. EDGE_TABLE_science-yixue.md contains only concept counts (Dantian 51 articles / Qigong 48 articles / Breathing 189 articles), lacking L1/L2/L3 levels and active/open/conflict judgments.

    Further Reading

    Evidence Level: C
    References: PMID 15014087, PMID 2362980, PMID 25098835, PMID 19899065, PMID 35721020, PMID 11901278, PMID 21639642, PMID 40779571, PMID 36644415


    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.

  • 同一份文件里两句话互相打架:不靠任何外部数据,也能判死一个衰减数值

    Evidence level: B

    先说结论

    一份讲”山体能不能挡住极低频电磁场”的文章里,有两句话是这么写的:

    • 正文说:1 km 厚花岗岩,衰减约 0.27 dB;
    • 同文表格说:100 m 厚花岗岩,衰减 27 dB。

    厚度相差 10 倍,衰减却相差 100 倍。而电磁波在导体里走,衰减和厚度是成正比的——厚 10 倍,衰减就该大 10 倍。所以这两句话里必有一句错,且不是”大概差一点”,是差了一千倍。

    关键不在于”算出了什么”,而在于”怎么发现的”:判死它,不需要查任何文献,不需要知道花岗岩的电导率到底该取多少,不需要任何地下实测。只需要这一份文件里的两个数字,和一个比例关系。

    这条检查现在已经写进审计工具(delta-table-scan.py v1.4 的 R5 项),全库自动跑。


    为什么这条检查值钱:它绕开了”参数争议”

    前面几轮审计一直在跟一个纠缠不清的问题打架:到底是谁错?

    δ/A 表里出现不自洽,有两种可能:

    1. δ 或 A 算错了(数值问题);
    2. σ 或 μr 本来就该取另一个值(参数问题)。

    第 2 种情况下,”不自洽”其实不是错——只是表格本意用了一个没写出来的等效参数。这就是前几轮反复出现”判了又推翻”的原因:第 12 轮曾判定”A 档 δ 被高估 37 倍”,第 13 轮自我证伪,发现真错在 C 档。

    R5 绕开了整个争议:它不比较”你写的值”和”我算的值”,它比较你自己写的两个值之间的比例。只要同一份文件里、同一材料、同一频点下出现两个”厚度–衰减”对,比例关系就必须成立——无论 σ 取多少。

    这是审计里很少见的一种判据:零外部依赖。它的误报率天然低,因为它不问”你对不对”,只问”你前后一致吗”。


    那道比例关系有多硬

    对于良导体中的平面波,趋肤深度

    $$\delta = \sqrt{\frac{2}{\omega \mu_0 \mu_r \sigma}}$$

    而厚度 $d$ 上的衰减(以 dB 计)近似为

    $$A_{dB} \approx 8.7 \cdot \frac{d}{\delta}$$

    把两式对各参数求偏导就能看出:在 δ 固定时,$A$ 对 $d$ 是严格一次齐次的,即

    $$\frac{A_1}{A_2} = \frac{d_1}{d_2}$$

    这个比例与 σ、μr、频率全部无关——它们都在分子分母里约掉了。这就是为什么这条检查不需要任何参数。

    顺便说一句,这个”一次齐次”性质也是本站另一篇(第 14 轮)关于”端点方向”的物理来源:因为 $A \propto \sqrt{\sigma}$,所以 σ 增大时 A 增大、δ 减小,两者方向相反。这次用的是它的厚度维度,上次用的是它的电导率维度——同一条式的两个侧面。


    实际跑出来的结果:全库只有 1 处

    工具在 780 全库(37 个文件、400 条数值断言)里找出 43 个可参与比例检验的点,从中构成 4 组”同材料、同频点、厚度至少差 2 倍”的配对,最终报出 1 处违反:

    项 内容
    文件 skin-depth-basement-building-mu-metal-table.md
    材料 / 频点 花岗岩 @ 7.83 Hz
    点 A L76:厚 100 m,A = 27 dB
    点 B L80:厚 1000 m,A = 0.27 dB
    厚度比 10×
    衰减比 100×
    偏差 10×(判据容差 2×)

    这正是第 13 轮靠人工对读才看见的那一处。当时是靠”散文 vs 表格”逐字对照发现的,属于工具盲区边缘;现在工具能自动复现,且只报这一处,没有顺带误报一片。

    一个反直觉的地方

    这份文件的立论是”山体对极低频几乎透明,所以靠山屏蔽不成立“。按复算(花岗岩 100 m ≈ 0.048 dB、玄武岩 100 m ≈ 1.53 dB),真实值比表里的 27 / 270 dB 更透明——也就是说,这批错把正确结论写得对它自己不利。

    所以按”看是否产生错误结论,不看倍数”的原则,不推翻任何结论,但数值必须改。这一点很重要:审计的价值不是”抓错显得厉害”,而是分清”错了但结论稳”和”错了且结论塌”。


    顺带修掉的一类更隐蔽的缺口:静默不入账

    做这条检查时撞上一个更麻烦的问题。

    工具的覆盖率一直是这么报的:”含 δ/A 列的表格 68 张,成功解析 45 张,未解析 23 张”。但实际情况是:有一批表既不算”已解析”,也不算”未解析”,它们根本没有进入分母。

    原因是表头写法。工具认 A(dB)、认 1 km A(dB),但不认 f=7.83 Hz (A, dB) 这种带逗号的写法。而全库唯一那处最严重的错表(就是上面那张),表头恰好就是这个写法。于是它:

    • 不进”已解析”(解析不出断言);
    • 也不进”未解析”(因为它连”这是 δ/A 表”都没被认出来,在分类那一步就被跳过了);
    • 报告里一个字都不提。

    这比”未解析”危险得多。”未解析”至少留下了一条可见的缺口;”静默不入账”连缺口都看不见。 修复后这类表被单独计数并列出(当前为 0),同时英文镜像里 $\delta$(7.83 Hz)、$\sigma$ (S/m) 这类 LaTeX 表头也一并纳入识别。

    修复的代价是:断言数 355 → 400(+45),表格分母 60 → 68,丢失 0 条。覆盖率从 68.3% 降到 66.2%——这不是退步,是因为分母里补进了原来被藏起来的表。覆盖率口径跨版本不可直接比较,这条规矩第 14 轮已经立过,这次又验证了一遍。


    修的时候自己造的三个假阳性(如实记)

    必须写下来,因为这是同一个失败模式的第 N 次复发:放宽匹配范围时没有同步收紧排除。

    # 我做了什么 造出了什么假阳性 怎么修的
    1 归一 LaTeX 表头后,$\\sigma\mu_r$ / $\\sigma\mu_r$ 这类比值列被认成 σ 列 取到 0.0996 这种”比值”当电导率,整行 δ 被判错 1 万倍 排除 implied / ratio / 比值 / 反推 / 推算
    2 散文中”抓第一个长度当厚度” 回看到了上一行表格里的厚度,或把 ASCII 图里标注的 δ 值当厚度 取消”第一个长度”启发式,改为三类显式绑定
    3 HAS_DB = r"dB" 无词边界 sandbox 里的 db 命中,一份无关文档被列为”含 dB 的表” 加词边界 (?<![A-Za-z])dB(?![A-Za-z])

    三个都不是读代码发现的,全部靠”① selftest 边界用例 ② 与上版基线逐条 diff 新出现的误报”。第 14 轮立的规矩”改审计工具前先留基线快照”,这一轮又救了三次。


    一个方法学副产品:判据”充分性”与判据”必要性”要分开

    这次还澄清了一件事,值得单独写下来。

    R5 这类内部一致性检查,只能判”不成立”,不能判”成立”:

    • 两处比例对得上 → 不能推出这两个数都对。它们可能用同一个错的 δ 算出来,错错相消。
    • 两处比例对不上 → 必然至少有一处错。

    本文那张 μ-metal 行就是活例子:它在 δ 表里被判”OK”,恰恰因为它用的是同一个错的 δ(0.57 cm)——§2 与 §3 错错相消。所以”通过 R5″不等于”没问题”,只等于”没自相矛盾”。

    这条区分对应审计里的一个老问题:必须把”证伪能力”和”验证能力”分开报。一个检查项如果只有证伪能力,就不能拿它的”通过数”当质量指标。


    证据与局限

    证据等级 E3(结构类比 + 库内文本核算),不是 E1。理由与边界:

    1. 全部为解析复算,无任何实测数据。本文 4 个 DOI 均为地物/材料类期刊(Studia Geophysica et Geodaetica、Geophysics、Water Resources Research、Clays and Clay Minerals),不在 PubMed 收录范围,故无 PMID;已逐条 Crossref 回查(标题/期刊/年份/作者全部一致)。
    2. “自洽值”不是”真值”。$A \approx 8.7d/\delta$ 与 $\delta=\sqrt{2/(\omega\mu\sigma)}$ 均假设均匀半空间、无位移电流、μr 与 σ 为常数。真实山体分层、各向异性、σ 随频率弥散(上述四篇均支持频散存在),所以任何”正确值”都只是在给定假设下的自洽值。
    3. “错”的判定范围很窄:仅指”同一文件内两处数字违反了 A ∝ d”。若原文本意是用某个未写出的等效 σ,问题性质从”算错”变为”没写清 σ”——这一点本文无法区分。
    4. 样本很小:43 个可判点里只构成 4 组有效配对,判据容差 2× 也是人为设定(选 2× 是因为它比观测到的 10× 偏差小一个量级,留了余量)。4 组配对不能外推为全库的自相矛盾率。
    5. 只覆盖等式断言。不等式断言(A ≥ 10 dB)不参与比例检验——因为不等式两端的最不利端语义会混进来,那是另一套判据(第 14 轮 R7/R8 已处理)。全库 400 条断言里有 23 条不等式,被本条检查有意排除。
    6. 工具只认 markdown 表格 + 特定散文句式。ASCII 图、纯叙述型描述仍在盲区;本文唯一那处发现是”表格 + 散文”跨形态对照,这恰好在工具能力的边缘。
    7. 覆盖率口径不可跨版本比:v1.3 的 68.3% 与 v1.4 的 66.2% 分母不同(补进了原来静默跳过的表),数字下降不代表质量下降。

    下一步

    1. R5 扩到”跨文件同材料”:目前只在单文件内互检。若两个文件对同一材料给出不同的”厚度–衰减”关系,同样违反 A ∝ d,但目前查不到。
    2. R5 与 R7 合并成一张”断言充分性矩阵”:R7 管端点方向(σ 区间),R5 管厚度比例,两者都是”零参数内部检查”,合并后可以统一报”这条断言的可判定性有几成”。
    3. 已发布正文的数值更正仍等 JOE 批(D1′/D1/D2/D7 一组),本文不改任何已发布正文。

    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。

    Two contradictory sentences in the same document: without any external data, you can still determine a decay value.

    Evidence level: B

    Conclusion

    A document discussing “whether a mountain can block extremely low-frequency electromagnetic fields” contains two contradictory statements:

    • The main text states: 1 km thick granite attenuates approximately 0.27 dB;
    • The same document’s table states: 100 m thick granite attenuates 27 dB.

    The thickness differs by a factor of 10, yet the attenuation differs by 100 times. For electromagnetic waves propagating in conductors, attenuation is proportional to thickness—increasing thickness by 10 times should increase attenuation by only 10 times. Therefore, one of these statements must be incorrect, and not merely “off by a little,” but off by a factor of 100.

    The key issue is not “what was calculated,” but “how it was discovered”: determining this inconsistency requires no external literature, no knowledge of the exact conductivity value for granite, and no underground measurements. It relies solely on two numbers within this single document and their proportional relationship.

    This check is now implemented in the audit tool (delta-table-scan.py v1.4, item R5) and runs automatically across the entire database.


    Why This Check Is Valuable: It Bypasses “Parameter Controversy”

    Previous audit rounds struggled with an unresolved question: Who is wrong?

    When the δ/A table shows self-inconsistency, there are two possibilities:

    1. δ or A was calculated incorrectly (numerical error);
    2. σ or μr should have taken a different value (parameter issue).

    In the second case, “self-inconsistency” is not necessarily an error—it may simply reflect that the table intended to use an equivalent parameter not explicitly stated. This explains why previous rounds repeatedly “judged and then overturned” results: Round 12 concluded that “A-grade δ was overestimated by 37 times,” but Round 13 self-refuted this, finding the real error lay in C-grade.

    R5 bypasses the entire controversy: it does not compare “your stated value” with “my calculated value.” Instead, it compares the ratio between two values you yourself stated. As long as the same document, same material, and same frequency point contain two “thickness–attenuation” pairs, the proportionality must hold—regardless of σ.

    This is a rare criterion in auditing: zero external dependency. Its false-positive rate is naturally low because it does not ask “are you right?” but only “are you consistent with yourself?”


    How Strong Is This Proportionality?

    For plane waves in good conductors, the skin depth is:

    $$\delta = \sqrt{\frac{2}{\omega \mu_0 \mu_r \sigma}}$$

    and the attenuation over thickness $d$ (in dB) is approximately:

    $$A_{dB} \approx 8.7 \cdot \frac{d}{\delta}$$

    Taking partial derivatives of both equations with respect to their parameters reveals: when δ is fixed, $A$ is strictly linearly homogeneous with respect to $d$, i.e.,

    $$\frac{A_1}{A_2} = \frac{d_1}{d_2}$$

    This ratio is independent of σ, μr, and frequency—they all cancel out in the numerator and denominator. This is why this check requires no parameters at all.

    Incidentally, this “linear homogeneity” property is also the physical origin of another article on this site (Round 14) concerning “endpoint direction”: because $A \propto \sqrt{\sigma}$, increasing σ increases A while decreasing δ, causing them to move in opposite directions. This time the check uses the thickness dimension of this relationship; previously it used the conductivity dimension—two complementary aspects of the same formula.


    Actual Results: Only One Violation Found

    The tool identified 43 points eligible for proportionality testing across the entire 780-document corpus (37 files, 400 numerical assertions). From these, it constructed 4 pairs of “same material, same frequency, thickness differing by at least 2×,” ultimately reporting 1 violation:

    Item Content
    File skin-depth-basement-building-mu-metal-table.md
    Material / Frequency Granite @ 7.83 Hz
    Point A L76: thickness 100 m, A = 27 dB
    Point B L80: thickness 1000 m, A = 0.27 dB
    Thickness ratio 10×
    Attenuation ratio 100×
    Deviation 10× (tolerance is 2×)

    This is precisely the one location detected manually in Round 13. At that time, it was found through careful word-by-word comparison of “prose vs. table,” lying at the edge of the tool’s blind spot; now the tool can automatically reproduce it, and reports only this single violation without any spurious false positives.

    A Counterintuitive Aspect

    The document’s argument is that “mountains are nearly transparent to extremely low frequencies, so shielding by mountains does not work.” Recalculating (granite 100 m ≈ 0.048 dB, basalt 100 m ≈ 1.53 dB), the true values are actually more transparent than the table’s 27 / 270 dB—meaning this batch of errors wrote the correct conclusion in a way that contradicts itself.

    Therefore, following the principle of “judging whether an error leads to a wrong conclusion, not just a wrong ratio,” no conclusions are overturned, but the numerical values must be corrected. This distinction is crucial: the value of auditing lies not in “catching errors to appear impressive,” but in distinguishing between “errors that do not affect the conclusion” and “errors that collapse the conclusion.”


    A More Subtle Gap Fixed: Silent Tables Excluded from Denominator

    During this check, a more problematic issue emerged.

    The tool’s coverage reporting has always been: “68 tables containing δ/A columns, 45 successfully parsed, 23 unparsed.” But in reality: a batch of tables is neither counted as “parsed” nor “unparsed”—they never entered the denominator at all.

    The reason lies in header formatting. The tool recognizes A(dB) and 1 km A(dB), but does not recognize headers like f=7.83 Hz (A, dB) that include commas. The most serious error table in the entire corpus (the one above) happens to use exactly this format. Consequently, it:

    • Does not enter “parsed” (cannot extract assertions);
    • Does not enter “unparsed” (because it wasn’t even recognized as a δ/A table during the classification step, so it was skipped entirely);
    • Appears nowhere in the report.

    This is far more dangerous than “unparsed.” “Unparsed” at least leaves a visible gap; “silent exclusion” hides the gap entirely. After fixing, such tables are now counted separately and listed (currently 0), and the English mirror now also recognizes LaTeX headers like $\delta$(7.83 Hz) and `$\sigma$ (S/m).

    The cost of this fix: assertion count increased from 355 to 400 (+45), table denominator from 60 to 68, with zero lost entries. Coverage dropped from 68.3% to 66.2%—this is not a regression, because the denominator now includes tables previously hidden. Coverage metrics across versions cannot be directly compared; this rule was established in Round 14 and verified again here.


    Three False Positives Created During Fixing (Recorded Honestly)

    Must document these, as this represents the Nth recurrence of the same failure mode: expanding match scope without tightening exclusion criteria.

    # What I Did What False Positive It Created How Fixed
    1 After normalizing LaTeX headers, ratio columns like $\sigma\mu_r$ / \sigma\mu_r$ were mistaken for σ columns Used 0.0996 such “ratios” as conductivity, causing δ to be judged wrong by 10,000× Exclude terms containing implied / ratio / 比值 / 反推 / 推算
    2 In prose, “grab the first length as thickness” Repeatedly saw thickness from the previous table row, or mistook δ values annotated in ASCII diagrams as thickness Cancel the “first length” heuristic, replacing it with three explicit binding types
    3 HAS_DB = r"dB" lacked word boundaries Matched db inside “sandbox,” causing an unrelated document to be flagged as “containing dB tables” Add word boundaries (?<![A-Za-z])dB(?![A-Za-z])

    None of these were discovered by reading code; all relied on “① self-test boundary cases ② diff against the baseline from the previous version to catch newly appearing false positives.” The rule established in Round 14—”save a baseline snapshot before modifying the audit tool”—saved us three times again this round.


    A Methodological Side Effect: Distinguish “Sufficiency” from “Necessity” of the Criterion

    This round also clarifies an important point worth documenting separately.

    R5 and similar internal consistency checks can only determine “does not hold,” not “holds”:

    • If two ratios match → cannot conclude both numbers are correct. They might both be derived from the same wrong δ, canceling each other’s errors.
    • If two ratios mismatch → necessarily at least one is wrong.

    The μ-metal row in this document exemplifies this: it passed R5 in the δ table precisely because it used the same wrong δ (0.57 cm)—§2 and §3 errors canceled each other out. Thus “passing R5” does not mean “no problem,” only “no self-contradiction.”

    This distinction corresponds to a longstanding issue in auditing: must report “falsification capability” and “verification capability” separately. If a check item has only falsification capability, its “pass count” cannot serve as a quality indicator.


    Evidence and Limitations

    Evidence level E3 (structural analogy + in-database text recalculation), not E1. Reasons and boundaries:

    1. All are parsed recalculations, with no measured data. The four DOIs cited in this article (Studia Geophysica et Geodaetica, Geophysics, Water Resources Research, Clays and Clay Minerals) are all earth/materials journals, not included in PubMed, hence no PMID; each was cross-checked via Crossref (title/journal/year/author all match).
    2. “Self-consistent values” are not “true values”. $A \approx 8.7d/\delta$ and $\delta=\sqrt{2/(\omega\mu\sigma)}$ both assume a uniform half-space, no displacement current, and constant μr and σ. Real mountains are layered, anisotropic, and σ exhibits frequency dispersion (all four cited papers support the existence of dispersion), so any “correct value” is only self-consistent under given assumptions.
    3. The scope of “wrong” is narrow: only refers to “two numbers within the same document violating A ∝ d.” If the original intent was to use an unstated equivalent σ, the nature of the problem shifts from “calculation error” to “unclear σ specification”—a distinction this article cannot resolve.
    4. Sample size is small: among 43 testable points, only 4 pairs formed valid matches; the tolerance of 2× is artificially set (chosen because it is one order of magnitude smaller than the observed 10× deviation, leaving margin). Four pairs cannot be extrapolated to represent the overall self-contradiction rate of the corpus.
    5. Only covers equality assertions. Inequality assertions (e.g., A ≥ 10 dB) do not participate in proportionality testing—because the worst-case semantics of both ends would mix, requiring a separate criterion (R7/R8 already handled this in Round 14). Among the 400 assertions in the corpus, 23 are inequalities and were intentionally excluded by this check.
    6. The tool only recognizes Markdown tables + specific prose patterns. ASCII diagrams and purely narrative descriptions remain blind spots; the single discovery in this article involved “table + prose” cross-format comparison, which lies exactly at the edge of the tool’s capability.
    7. Coverage metrics cannot be compared across versions: v1.3’s 68.3% and v1.4’s 66.2% have different denominators (the latter includes tables previously silently skipped), so the numerical drop does not indicate quality decline.

    Next Steps

    1. Expand R5 to “cross-file same material”: currently only checks within single files. If two files assign different “thickness–attenuation” relationships for the same material, this also violates A ∝ d, but cannot yet be detected.
    2. Merge R5 and R7 into a single “assertion sufficiency matrix”: R7 manages endpoint direction (σ intervals), R5 manages thickness ratios; both are “zero-parameter internal checks.” Merging them will uniformly report “to what extent is this assertion determinable.”
    3. Numerical corrections in published prose await JOE approval (D1′/D1/D2/D7 group); this article makes no changes to already-published prose.

    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.

  • 静息态其实不是静息:三态阶梯里「知道在被测量」占掉多少,以及 Δ 该怎么重算

    Evidence level: C

    一个人「随便躺着」时的腹式占比,只要身上绑着呼吸带,就已经被动过一次——只是没人记录动了多少。

    一、Δ 是个减法,分子干净、分母不干净

    「练功 vs 不练功」最后都落在一个减法上:Δ = P 态减 R 态。这个写法默认被减的那个数不受干预。可 R 态从来不是真空:受试者知道身上有带子、知道有人在看数据、知道这一节叫「静息」。呼吸恰好是少数既能自动跑、又能被随手接管的功能。于是同一个 Δ 有两种读法:「可调用的技巧储备」,或「控制强度的差值」。两者符号相同、方向一致,光比大小分不开;要分开,只能把 R 态本身拆开量。

    二、注意呼吸与刻意呼吸走两套部分重叠的线

    有直接成像证据:同一个呼吸动作被「刻意执行」与被「注意」时,动用的皮层网络不同但相互重叠,随意控制通道与注意调制通道可分别激活(PMID 28954895)。所以「我只是躺着、只是注意到呼吸」和「我在做腹式呼吸」不是同一件事的两端,而是两条能各自拉起的线。更麻烦的是时间尺度——内感受注意随呼吸周期本身调制脑-心耦合(PMID 35964864),污染量不是常数,一个呼吸周期内就上下摆。

    三、分心不是「回到基线」,是另一种偏离

    最直觉的补救是加一个分心态(R+D):让人做连续减 7 或 n-back,注意力占满就回到无控制了。但它有硬证据支持「分心确实改变呼吸」,同时否掉了「改变方向就是回到基线」。认知负荷对呼吸的系统综述给出分化图景:精神负荷的标志是呼吸变快、分钟通气量升高,幅度基本稳定;呼吸率总变异性不受影响,但相关成分下降;潮气末 CO2 下降,存在过度通气倾向(PMID 27403347)。分心态偏频率与气体交换、不偏幅度。这就堵死了推理:R+D 相对 R 往哪边动,取决于「注意占满」与「负荷致通气增强」谁大,这个相对大小文献没有数——必须实测。

    四、分心还会改掉呼吸与心搏的耦合本身

    有研究用虚拟现实 Go-NoGo 诱导两级认知负荷,同步记 RR 间期、收缩压与呼吸:负荷升高时 RRI-SBP 与 RRI-呼吸的相互作用都下降,SBP-呼吸的线性相关未变、互信息反而升高,而带余弦调制双指数拟合后差异最显著的参数恰在这一对之间(PMID 39260046)。含义是:分心不只让呼吸换个样子,它改掉了呼吸与其他信号的耦合结构。而 φ 与 %AB 正是从胸腹两条带的相对时序算出的量——耦合一变,这个时序的可解释性就变了。

    五、φ 的噪声底比判据那两条线还宽

    任何「差多少才算真差」的判据都得先跟噪声底比。φ 这一侧有明确记录:早产儿安静睡眠中胸腹相位角的个体内变异系数达 11%–123%,作者用机械模型排除后判定变异来自胸腹相位关系本身的不稳、而非算法(PMID 24716708)。成人侧的问题不是噪声大小而是没人报得齐:31 项研究(1994–2025)的范围综述指出信度指标报告不统一、绝对容积一致性随校准与体位漂移、没有单一模态够格临床可用(PMID 42862511)。「φ 差 > 5 度」这类阈值的可靠性因此取决于设备、体位与时长,而既往研究几乎不重样。

    六、三格判据,但边界得先标定

    第一格(污染成立):R+D 与 R 的 %AB 差 > 3 个百分点或 φ 差 > 5 度,且方向朝「更接近新手基线」,则 R 态确认含隐性控制,Δ 改用 P − (R+D) 重算,两态旧设计降级为含反应性偏差的上界估计。第二格(可忽略):两态无差异,两态设计成立。第三格(非单调):分心反使 %AB 上升,说明负荷与呼吸的关系非单调,三态不够、须加认知负荷维度。但 3pp 与 5° 都没有出处,是照写作习惯取的整数;正确做法是换成同一批人隔日重测算出的 MDC95 倍数——先定噪声再定阈值。用未标定的阈值判污染,等于用没校过的尺子量长度。

    七、测法,与一个必须承认的缺陷

    ≥36 人,新手 / 1–5 年 / 5 年以上各 12;同一时段按拉丁方平衡顺序测三态——R(仅告知躺着休息)、R+D(无呼吸提示 + 连续减 7 或 n-back,同记任务成绩作操纵检验)、P(明确指令做丹田式腹式呼吸);每态 ≥10 分钟,取稳定段算 %AB、φ(互相关法)、呼吸频率、潮气量与潮气末 CO2。加 30 秒窗滑动看时间进程——反应性若存在通常随熟悉衰减,这条曲线本身即证据。

    必须承认:三态只覆盖「指令强度」一个轴上的三个点,真正的零档(不知晓自己呼吸被记录)是缺的。缺了它,阶梯只给斜率、截距悬空。

    L1 网络状态:该主题尚未进入 L1 对应网络。 CORRESPONDENCE_NETWORK.md 节点表按「易学/中医侧」「物理侧」「生理/体验侧」三类域分派,生理侧八项(HRV / 皮温皮电 / 呼吸率 / 呼吸模式 / EMG / EEG / 胃肠蠕动 / 气感报告 / 体质分型)全为横断面观测项,其中「呼吸率 / 呼吸模式 | 可穿戴 | 与节律耦合最直接」仅登记「观测什么」,不含任何「知晓 / 遮蔽 / 反应性 / 指令强度」维度;边表 3.1–3.5 只登记五行↔五脏/五色/五音/五方/五季/五志、八卦↔五行、天干地支↔五行、六十甲子纳音↔频率,止于 E-050(「六十甲子纳音 | 对应 | 60 个频率」,层级 L1、状态 open,现代证据栏原文「尚无任何可测来源」),无任何以测量反应性、遮蔽或指令强度为源或目标的边。故本文证据等级封顶 C,正文所有结论均不构成已验证机制。

    证据与局限

    已验证:注意呼吸与刻意控制呼吸动用不同但重叠的皮层网络(PMID 28954895);内感受注意随呼吸周期调制脑-心耦合(PMID 35964864);认知负荷使呼吸变快、分钟通气量升高、幅度稳定、相关成分下降、潮气末 CO2 下降(PMID 27403347);认知负荷改变 RRI-呼吸与 RRI-SBP 耦合结构(PMID 39260046);φ 个体内变异系数 11%–123% 且源于生理不稳(PMID 24716708);成人呼吸模式测量信度报告不统一、无单一模态临床可用(PMID 42862511)。

    部分验证:认知分心能否把呼吸拉回「无控制基线」——文献方向不一致,第三节已说明两股力的相对大小未知,这正是本课题要测的。

    待验证:① 3pp / 5° 两条线的实证位置(须用 MDC95 标定);② 反应性的时间衰减曲线形态;③ 真正的遮蔽零档与三态阶梯之间的落差;④ 年限越长是否越容易在 R 态自动进入控制(若成立,年限斜率会被压成假阴性)。

    相关阅读

    证据等级: C

    参考文献: PMID 28954895, PMID 35964864, PMID 27403347, PMID 39260046, PMID 24716708, PMID 42862511


    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。

    Resting State is Not Actually Resting: How Much \”Awareness of Being Measured\” Occupies the Three-State Ladder and How to Recalculate Δ

    Evidence level: C

    When a person lies “casually,” the abdominal breathing proportion is already passively perturbed simply by wearing a breathing belt, even if no one records how much they moved.

    I. Δ Is a Subtraction Where the Numerator Is Clean but the Denominator Is Not

    The comparison between “practicing” and “not practicing” ultimately reduces to a single subtraction: Δ = P-state minus R-state. This notation implicitly assumes the subtracted number (R) is unperturbed. However, the R-state is never a vacuum: subjects know they have a belt on their body, know someone is watching the data, and know this section is called “resting.” Breathing is precisely the rare function that can run automatically yet be easily hijacked. Thus, the same Δ has two readings: “reservable skill stock” or “difference in control intensity.” They share the same sign and direction; magnitude alone cannot separate them. To distinguish them, one must decompose the R-state itself for measurement.

    II. Attention and Intentional Breathing Follow Two Partially Overlapping Lines

    Direct imaging evidence exists: the same breathing action engages different but overlapping cortical networks when “intentionally executed” versus when “attended to” (PMID 28954895). Consequently, “I am just lying down, just noticing my breathing” and “I am doing abdominal breathing” are not two ends of the same spectrum but rather two distinct lines that can each be pulled. More complicating is the timescale: interoceptive attention modulates brain-heart coupling over the breathing cycle itself (PMID 35964864). Therefore, the perturbation amount is not constant; it fluctuates up and down within a single breathing cycle.

    III. Distraction Is Not “Returning to Baseline” but Another Deviation

    The most intuitive remedy is adding a distraction state (R+D): have subjects perform continuous subtraction of 7 or n-back tasks until attention is fully occupied, theoretically returning to an uncontrolled baseline. However, hard evidence supports that “distraction indeed alters breathing,” while simultaneously refuting the claim that “changing direction equals returning to baseline.” A systematic review of cognitive load on breathing presents a differentiated picture: markers of mental load include faster breathing, increased minute ventilation, stable amplitude, decreased correlated components, and reduced end-tidal CO2, indicating a tendency toward hyperventilation (PMID 27403347). Distraction biases frequency and gas exchange, not amplitude. This blocks the reasoning: the direction of movement of R+D relative to R depends on whether “attention saturation” or “load-induced ventilation enhancement” is stronger—a relative magnitude that literature has not quantified; it must be measured empirically.

    IV. Distraction Also Alters the Coupling Structure Between Breathing and Heartbeat

    Research using virtual reality Go-NoGo tasks induced two levels of cognitive load, synchronously recording RR intervals, systolic blood pressure (SBP), and breathing. When load increased, the interactions between RRI-SBP and RRI-breathing both decreased; the linear correlation between SBP-breathing remained unchanged, yet mutual information actually increased. After fitting with cosine-modulated double-exponential models, the most significant difference parameter lay precisely between this pair (PMID 39260046). The implication is clear: distraction does not merely change breathing’s appearance; it alters the coupling structure between breathing and other signals. Since φ and %AB are quantities derived from the relative timing of chest and abdominal bands, once this coupling changes, the interpretability of this timing shifts accordingly.

    V. The Noise Floor of φ Is Wider Than the Two Criteria Lines

    Any criterion for “how much difference counts as true difference” must first be compared against the noise floor. On the φ side, clear records exist: in premature infants during quiet sleep, the intrasubject coefficient of variation of the chest-abdominal phase angle ranged from 11% to 123%; after excluding mechanical models, the authors concluded the variation stemmed from instability in the chest-abdominal phase relationship itself, not the algorithm (PMID 24716708). The issue on the adult side is not noise magnitude but inconsistent reporting: a scope review of 31 studies (1994–2025) indicates non-uniform reliability metrics, absolute volume consistency drifting with calibration and posture, and no single modality qualifying for clinical use (PMID 42862511). Therefore, the reliability of thresholds like “φ difference > 5 degrees” depends on equipment, posture, and duration—factors that prior studies rarely standardized.

    VI. Three-Grid Criteria, But Boundaries Must Be Calibrated First

    First grid (pollution established): If the difference in %AB between R+D and R exceeds 3 percentage points, or φ difference exceeds 5 degrees, and the direction moves toward “closer to novice baseline,” then the R-state is confirmed to contain hidden control. Δ must be recalculated as P − (R+D). The original two-state design degrades into an upper-bound estimate containing reactive bias.

    Second grid (negligible): If there is no difference between the two states, the two-state design holds.

    Third grid (non-monotonic): If distraction causes %AB to rise, it indicates the relationship between load and breathing is non-monotonic; three states are insufficient, and a cognitive load dimension must be added.

    However, the 3pp and 5° thresholds lack provenance; they are merely integers chosen by writing convention. The correct approach is to replace them with MDC95 multiples derived from re-measuring the same cohort on alternate days—first calibrate the noise, then set the threshold. Using uncalibrated thresholds to judge pollution is like measuring length with an uncalibrated ruler.

    VII. Measurement Method and One Admitted Defect

    ≥36 participants: 12 novices, 1–5 years experience, and >5 years experience. Within the same time window, three states were measured using Latin square balanced order:
    – R-state: Instructed only to lie down and rest.
    – R+D-state: No breathing prompts + continuous subtraction of 7 or n-back; task performance recorded as manipulation check.
    – P-state: Explicit instruction to perform Dan Tian-style abdominal breathing.

    Each state lasted ≥10 minutes; stable segments were used to calculate %AB, φ (via mutual correlation method), breathing rate, tidal volume, and end-tidal CO2. A 30-second sliding window examined temporal progression—if reactivity exists, it typically decays with familiarity; this decay curve itself constitutes evidence.

    Admitted limitation: The three states cover only three points along the “instruction intensity” axis. The true zero level (unaware that breathing is being recorded) is missing. Without it, the ladder provides only slope and intercepts remain suspended.

    L1 Network State: This topic has not yet entered the L1 corresponding network. The CORRESPONDENCE_NETWORK.md node table assigns across three domains: “Yijing Studies/TCM side,” “Physical side,” and “Physiological/Experiential side.” The physiological side includes eight items (HRV / skin temperature/skin conductance / breathing rate / breathing pattern / EMG / EEG / gastrointestinal motility / qi sensation report / constitution typing), all cross-sectional observations. Among them, “breathing rate / breathing pattern | wearable | most directly coupled with rhythm” registers only “what is observed,” containing no dimensions of “awareness / masking / reactivity / instruction intensity.” Edge tables 3.1–3.5 list only Five Phases ↔ Five Organs/Five Colors/Five Tones/Five Directions/Five Seasons/Five Emotions, Eight Trigrams ↔ Five Phases, Heavenly Stems/Earthly Branches ↔ Five Phases, and Sixty Jiazi Na Yin ↔ Frequency, stopping at E-050 (“Sixty Jiazi Na Yin | Correspondence | 60 Frequencies,” level L1, state open, original text in modern evidence column states “no measurable source yet“), with no edges sourced or targeted by measurement of reactivity, masking, or instruction intensity. Therefore, this article’s evidence level caps at C; all conclusions in the main text do not constitute verified mechanisms.

    Evidence and Limitations

    Verified: Attention to breathing and intentional control of breathing engage different but overlapping cortical networks (PMID 28954895); interoceptive attention modulates brain-heart coupling over the breathing cycle (PMID 35964864); cognitive load increases breathing rate, elevates minute ventilation, stabilizes amplitude, reduces correlated components, and lowers end-tidal CO2 (PMID 27403347); cognitive load alters the coupling structure between RRI-breathing and RRI-SBP (PMID 39260046); φ intrasubject coefficient of variation ranges from 11% to 123% and stems from physiological instability (PMID 24716708); adult breathing pattern measurement reliability reports are inconsistent, with no single modality clinically viable (PMID 42862511).

    Partially Verified: Whether cognitive distraction can pull breathing back to an “uncontrolled baseline”—literature directions are inconsistent; Section III already explained the unknown relative magnitude of these two forces, which is precisely what this study must measure.

    To Be Verified: ① Empirical positions of the 3pp / 5° lines (must be calibrated via MDC95); ② Shape of the reactivity time-decay curve; ③ Gap between the true masking zero level and the three-state ladder; ④ Whether longer experience makes it easier to automatically enter control in the R-state (if true, the experience slope would be suppressed into false negatives).

    Related Reading

    Evidence Level: C

    References: PMID 28954895, PMID 35964864, PMID 27403347, PMID 39260046, PMID 24716708, PMID 42862511


    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.

  • 同一锅汤,换个下料顺序就换个结论:独立假设与相关假设下谐波方差份额的偏差

    Evidence level: C

    “几何占四成、弹性占三成”——这句话漏了一个状语:在假定三者互不相关的前提下。换掉这个前提,四成可能变两成。

    正文

    一、一个数字少了状语,就等于没有数字

    前几篇把脉象谐波的能量分配拆成”几何 / 弹性 / 边界条件”三份账,并且已经点出一个坑:分解顺序会改变各家的份额。但那一步只到方法学处方为止——改用 Shapley 效应、分层报告、附上差值——始终没有回答”差多少”。

    而”差多少”恰恰是这个课题能不能继续的前提。因为方差基分解有一个几乎从不被写出来的前提:输入之间统计独立。前提成立,”占几成”是一个确定的数;前提不成立,同一个模型、同一批人、同一个采样路径,能算出好几个都”正确”的份额。此时任何单设定份额都不可比、不可复现,也不能拿去和别人的数字对照。

    二、独立假设在建模圈里是常态,而常态不等于成立

    这个假设的普遍程度有书目计量证据:过去十年里绝大多数已发表的敏感性分析仍是局部法或单因子轮换(OAT),并且建立在未经检验的线性与可加性假设上,全局方法只有少数人在用(PMID 26934843)。

    更直接的一条来自 2026 年的方法学文章:传统方差基全局敏感性分析依赖输入独立的假设,而这个假设对经贝叶斯校准的模型几乎从不成立;近年的 Rosenblatt 变换与 Shapley 效应扩展理论上解决了这一限制,但实施时需要对相关联合分布做条件采样,而现有方案对输入依赖结构仍带限制性假设(PMID 41838759)。

    换句话说:不是没人知道该怎么做,是做起来贵,所以默认不做。

    三、脉象这条线上,”输入相关”不是可能性,是实测

    如果三个输入真的互不相干,独立假设只是省事;但这条线上它们明显同向走动。

    谐波失真(HD,基频以上能量与基频能量之比)在 24 小时连续遥测里随收缩压线性下降,同时与切向模量、脉搏波速度线性相关,且两个饮食组的斜率显著不同(PMID 35433650)——一个指标同时随血压、硬度、时间走。

    传输线模型那一侧更直接:增强指数(AI)、脉搏波速度(PWV)、谐波失真三者都主要对结构刚度敏感,而 HD 与 AI 还与几何参数强相关(PMID 40097550)。三个”独立的解释变量”,被同一个观测量同时读取。

    四、更麻烦的一层:弹性本身不是常数

    份额分解默认每个参数是一个人身上的固定值。但血管壁的相对粘性在同一个下午就能被改:抑制一氧化氮合成后从 34.9±8.9% 降到 28.9±8.6%,阻断钙激活钾通道后从 31.7±6.6% 升到 35.7±8.0%,两者合用升幅更大(PMID 29158357)。

    这不是”个体差异”,是急性调控。把一个受内皮调控的量当成常量塞进方差分解,等于把一部分本属”状态”的方差强行记到”参数”名下——份额自然被抬高。

    五、这件事已经被正面做过一次,说明可行

    生物力学领域已有现成路径:对典型的脉搏波传播模型做考虑输入相关性的全局敏感性分析,用代理模型把计算成本压到理论上的 1/27000,结论明确写着”输入相关性会显著影响敏感性分析结果,因此在模型开发阶段必须彻底调查输入相关性的影响”(PMID 38116742)。

    技术门槛已经不高了。真正缺的是一句判据:差值到多大才算不能忽略。

    六、把”差值”本身做成一把尺子

    这里给一个可执行的判据,也是本文最想提出的一点:不要再问”几何占几成”,改问”在什么相关性结构下占几成“,并把两种设定之差当作”共线强度”的定量指标直接发表。

    • 差值 < 5 个百分点:共线弱,既往独立假设下的结论可保留,只需在方法里注明;
    • 差值 5–15 个百分点:必须双设定并列报告,任何单设定份额不得单独引用;
    • 差值 > 15 个百分点(尤其在 C5 以上高阶):既往基于独立假设的份额结论一律作废、须按相关输入重算。

    七、怎么测

    ≥60 人的个体化一维血流模型,用同一条生理范围采样路径,跑两遍:① 各参数按边缘分布独立抽样;② 按实测的人群几何—弹性—边界相关性构造联合分布(或用后验相关)。两遍各算逐阶谐波能量的 Shapley 方差分解,分离主效应与交互项,逐阶比对份额差。

    注意证据薄弱:以上全部是方法学推演加间接证据。L1 对应网络里没有这条边,一维模型份额能否外推到真实脉搏尚无直接验证,本结论只能当作”份额数字必须带假设标签”这一操作规范的论证,不能当作”几何占几成”的答案。

    证据与局限

    已验证:输入相关性会显著改变生物力学模型的敏感性分析结果,且代理模型可把相关输入的全局分析成本降到 1/27000(PMID 38116742);脉象侧三个常用指标对刚度、几何、血压存在重叠依赖(PMID 35433650、PMID 40097550);血管壁粘性受急性内皮调控而非常数(PMID 29158357)。

    部分验证:独立输入假设在建模实践中普遍存在且常未被检验(PMID 26934843、PMID 41838759)——这是”问题存在”的证据,不是”偏差有多大”的证据;5 / 15 个百分点两条线目前无实证出处,是暂定阈值。

    待验证:两设定份额差在真实脉象上究竟落在哪一档;逐阶差值是否随阶次单调放大;共线强度能否作为可复现的个体特征。

    相关阅读

    证据等级: C
    参考文献: PMID 38116742, PMID 41838759, PMID 26934843, PMID 35433650, PMID 40097550, PMID 29158357
    L1 网络状态:「节点表按『易学/中医侧』『物理侧』『生理/体验侧』三类域分派;物理侧仅列『相位 / 节律 —— 周期性耦合的关键』,生理侧八项全为可穿戴单点横断面观测」——该主题(谐波方差份额的输入相关性 / 共线强度)尚未进入 L1 对应网络,边表止于 E-050「六十甲子纳音 | 对应 | 60 个频率」,现代证据栏原文「尚无任何可测来源」。


    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。

    Same Soup, Different Conclusion: Bias in Harmonic Variance Share Under Independent versus Correlated Assumptions

    Evidence level: C

    “Geometry accounts for 40%, elasticity for 30%” — this statement omits a crucial adverbial clause: under the assumption that the three factors are mutually uncorrelated. Altering this premise could shift the 40% figure to 20%.

    Main Text

    I. A missing adverbial clause renders a number meaningless

    Previous articles decomposed the energy allocation of pulse wave harmonics into three accounts: “geometry / elasticity / boundary conditions,” and highlighted a pitfall: the decomposition order alters each party’s share. That step stopped at the methodological prescription — using Shapley effects, layered reporting, and appending deltas — without ever answering “by how much”.

    Yet “by how much” is precisely the prerequisite for continuing this inquiry. Variance-based decomposition rests on an almost never-stated premise: statistical independence of inputs. If this premise holds, “what share” is a determined number; if it fails, the same model, same cohort, and same sampling path can yield multiple “correct” shares. Under such conditions, any single-assigned share becomes incomparable, irreproducible, and unusable for cross-validation against others’ numbers.

    II. The independent assumption is standard in modeling circles, but standard does not mean valid

    The ubiquity of this assumption is evidenced by bibliometrics: over the past decade, most published sensitivity analyses remain local methods or single-factor rotation (OAT), and rely on untested linearity and additivity assumptions, while global methods are used by only a minority (PMID 26934843).

    A more direct source comes from a 2026 methodological article: traditional variance-based global sensitivity analysis depends on the input independence assumption, which rarely holds for Bayesian-calibrated models; recent Rosenblatt transformations and Shapley effect extensions theoretically resolve this limitation, but implementation requires conditional sampling of correlated joint distributions, while existing schemes still impose restrictive assumptions on input dependency structures (PMID 41838759).

    In other words: it is not that no one knows how to proceed, but it is expensive to do so, so it is defaulted out.

    III. On the pulse wave line, “input correlation” is not a possibility but an empirical fact

    If the three inputs were truly uncorrelated, the independent assumption would merely be a convenience; but along this line they clearly move in tandem.

    Harmonic distortion (HD, the ratio of energy above fundamental to fundamental energy) linearly decreases over 24-hour continuous telemetry with systolic pressure, while simultaneously showing linear correlation with tangential modulus and pulse wave velocity (PWV), with significant differences in slopes between the two dietary groups (PMID 35433650) — one indicator thus varying with blood pressure, stiffness, and time.

    The transmission line model side is even more direct: augmentation index (AI), PWV, and HD all primarily depend on structural stiffness, while HD and AI also strongly correlate with geometric parameters (PMID 40097550). Three “independent explanatory variables” are simultaneously read by the same observed quantity.

    IV. A further complication: elasticity itself is not constant

    Share decomposition defaults to each parameter being a fixed value for an individual. Yet the relative viscosity of the vessel wall can be altered within the same afternoon: inhibiting nitric oxide synthesis reduces it from 34.9±8.9% to 28.9±8.6%, blocking calcium-activated potassium channels raises it from 31.7±6.6% to 35.7±8.0%, and combining both yields an even larger increase (PMID 29158357).

    This is not “inter-individual variation,” but acute regulation. Plugging a quantity governed by endothelial control into variance decomposition as a constant equals forcing part of the variance inherently belonging to “state” onto the account of “parameter” — naturally inflating the share.

    V. This has been positively addressed once, proving feasibility

    The biomechanics field already possesses a ready path: performing global sensitivity analysis considering input correlation on typical pulse wave propagation models, using surrogate models to reduce computational cost to theoretically 1/27000, with conclusions explicitly stating “input correlation significantly affects sensitivity analysis results, so input correlation effects must be thoroughly investigated during model development” (PMID 38116742).

    The technical threshold is now low. What truly lacks is a criterion: to what extent must the difference be considered non-negligible.

    VI. Turning “difference” itself into a ruler

    Here we provide an actionable criterion, which is the most important point of this paper: stop asking “what share does geometry hold?” and instead ask “under what correlation structure does it hold,” and publish the difference between the two settings as a quantitative indicator of “collinearity strength.”

    • Difference < 5 percentage points: weak collinearity; conclusions under previous independent assumptions may be retained, provided the method notes this;
    • Difference 5–15 percentage points: both settings must be reported side-by-side; no single-set share may be cited alone;
    • Difference > 15 percentage points (especially at C5 and higher orders): all previous share conclusions based on independent assumptions are void and must be recalculated using correlated inputs.

    VII. How to measure

    ≥60 individuals’ one-dimensional blood flow models, sampled along the same physiological range path, run twice: ① each parameter independently sampled from marginal distributions; ② joint distribution constructed from measured population geometry-elasticity-boundary correlations (or posterior correlation). For each run, compute Shapley variance decomposition of harmonic energies, separating main effects and interaction terms, then compare share differences order-by-order.

    Note on weak evidence: All above are methodological deduction plus indirect evidence. L1 network lacks this edge; whether one-dimensional model shares can be extrapolated to real pulse waves remains unverified directly. This conclusion serves only as argument for the operational norm that “share numbers must carry assumption labels,” not as an answer to “what share does geometry hold.”

    Evidence and Limitations

    Verified: Input correlation significantly alters biomechanical model sensitivity analysis results, and surrogate models can reduce global analysis cost of correlated inputs to 1/27000 (PMID 38116742); three common pulse wave indicators exhibit overlapping dependence on stiffness, geometry, and blood pressure (PMID 35433650, PMID 40097550); vessel wall viscosity is acutely regulated by endothelium and not constant (PMID 29158357).

    Partially verified: Independent input assumption is widespread in modeling practice and often untested (PMID 26934843, PMID 41838759) — this is evidence that “the problem exists,” not evidence of “how large the deviation is”; the 5/15 percentage point thresholds currently lack empirical basis and are provisional.

    To be verified: Where exactly the difference between two settings falls on real pulse waves; whether incremental differences monotonically increase with order; whether collinearity strength can serve as a reproducible individual characteristic.

    Related Reading

    Evidence level: C
    References: PMID 38116742, PMID 41838759, PMID 26934843, PMID 35433650, PMID 40097550, PMID 29158357
    L1 network status: Nodes assigned across three domains — “Yijing Studies / TCM side,” “physics side,” and “physiological/experience side”; the physics side lists only “phase / rhythm — key periodic coupling,” while the physiological side includes eight items, all cross-sectional wearable observations. This topic (harmonic variance share input correlation / collinearity strength) has not yet entered the L1 network; the edge table stops at E-050 “Sixty Jiazi Na Yin | Correspondence | 60 frequencies,” with the modern evidence section originally stating “no measurable source yet.”


    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.

  • 把账做平不等于账对了:残差是「调个参数」能救,还是模型少了一个维度

    Evidence level: C

    残差能被某个参数吃掉,不代表这个参数是对的——它可能只是在替一个还不存在的维度打工。

    正文

    一、同一个「对上了」,有两条来路

    一段脉搏波形拟合不好,剩下一截残差。摆在面前只有两条路:要么某个参数取值不对,调准就行;要么模型本身就少了一个维度,怎么调也调不到。前者是小修,后者是大修。

    这两条路通向相反的结论。若只是参数没个体化,共振腔那条假说还站着;若缺的是结构性维度,那么在补上结构项之前,「几何占几成、弹性占几成」这类份额数字全部悬空。麻烦在于两条路表面一模一样,都是「调完曲线更贴了」。分开它们靠的是换一套判据。

    二、可辨识性本来就分两层

    心血管药理建模里有个干净案例:一个非线性血流动力学反馈模型总有参数估不准,结构可辨识性分析确认模型本身不可辨识,重参数化(削减参数表)后才变成全局可辨识。关键在于,重参数化前后的两个模型在输入-输出行为上完全无法区分(PMID 22343490)。这句话很重:只比对曲线,永远分不出「参数多的版本」和「参数少的版本」哪个是真的。

    Profile likelihood 这条路线把两层拆得更干脆:结构不可辨识表现为功能相关的参数组合——换一组参数给出同样输出;实践不可辨识则来自数据量和质量不够(PMID 19505944)。前者加样本救不回来,后者能救。所以「残差压不下去」必须先分两层问,否则后面全是猜。

    三、实践层有硬数字

    闭环集总参数模型的系统性测试给了一组能直接用的参照:无噪声时所有参数都能标到万分之一以下误差;只要加上 5% 正态噪声,最敏感那五个参数误差被限在 10% 上下,最不敏感那四个用波形数据无法可靠估计,最不敏感那三个「特别难估」——作者的结论是优先去测它们,而不是去调它们(PMID 34758345)。同一份研究还发现,用压力波形做代价函数,比用每搏量、脉压这类临床指标准几个数量级。

    这三个数字给了一条能搬的判据:残差压不到噪声底,先看这个参数在敏感性排序里排第几。末位那几个压不下去是正常的,属于「没测够」,不构成「模型错了」的证据。

    四、结构性错误的签名是方向不对

    非线性弹性管的实验给出了一个很干净的签名。同一根管、两种生理脉幅:线性方法在高阶散射明显增大,而且随脉幅升高,三点法在低阶给出物理上不成立的负衰减(PMID 10468357)。负的衰减系数没有任何物理解释——那是模型形式的错,不是参数值的错。

    这就是两种错误的分水岭。参数错,残差有界、方向大体随机;结构错,残差会越界到物理不可能的象限。所以判据第一问不该是「残差多大」,而是「它有没有跑到不可能的象限里」。

    五、反过来更危险:越能对上,越可能在跨栏补偿

    谐波失真在 24 小时内随收缩压升高线性下降(PMID 35433650),说明幅值依赖是一个独立维度;模型里没有它,它就会以「某个参数随血压漂移」的形式冒出来。而交感能在独立于血流动力学的条件下调节大动脉硬度(PMID 33035441),等于多配了一个自由旋钮,专门用来与别的参数互相抵消。更底层:小动脉在肌源性与代谢调节受限时,局部阻力由血液表观粘度而非管径决定,表观粘度本身随灌注压与血流改变(PMID 26011432)——「阻力」在生理上本就不是常数。

    所以「参数被调准了」必须附加条件:它在什么生理状态下准、能准多久。一个几何参数若随血压做小时级漂移,它不是被调准了,是在替缺失的幅值依赖维度打工。相关输入下的多解与参数补偿还会让份额估计进一步失真(PMID 41838759)。

    六、判据三格

    第一格,结构层先判:重参数化后输入-输出不可区分,或残差落进物理不可能象限(负衰减、非单调)→ 判结构性缺维度,属大修。份额数字一律作废,须先补压力依赖、频率依赖粘性这类结构项再谈分账。

    第二格,实践层:单参数扫描能把该阶残差压到重复测量噪声底,且压完后参数值在隔日重复中稳定 → 判参数未个体化,属小修,假说保留。

    第三格最易误判:残差压到噪声底,但参数值随瞬时血压或心率显著漂移 → 判吸收成功但归因错误。这一格必须单列,因为它表面「修好了」,实质是把结构缺失的方差记到了假参数名下。

    七、测法

    ≥30 名受试者,每人隔日重复 3 次,估计每阶残差的噪声底与参数值的重测稳定性。对每一阶残差依次做三件事:单参数扫描(能否压到噪声底)、多参数联合反解(解是否沿脊线漂移、是否唯一)、残差与瞬时血压心率的逐阶相关(幅值依赖的直接读数)。敏感性排序先算出来,把末位参数预先归入「没测够」一类,避免污染「模型错了」的判断。

    一句话收尾:残差是一张收据,它告诉你账没平,但不告诉你钱花在哪儿。要判小修还是大修,得看参数在什么条件下才准,而不是它准的时候有多准。

    证据与局限

    已验证:结构可辨识性与实践可辨识性必须在方法上分开处理,且重参数化前后的模型可以在输入-输出行为上完全无法区分(PMID 22343490、PMID 19505944);带 5% 噪声时,心血管集总参数模型中最不敏感的几个参数确实无法靠波形数据可靠估计(PMID 34758345);非线性弹性管在高脉幅下线性方法会给出物理不成立的负衰减(PMID 10468357);动脉硬度可被交感独立调节、表观粘度由灌注压与血流决定(PMID 33035441、PMID 26011432)。这些是「判别该怎么做」的方法学证据,强度较高。

    部分验证:「残差压到噪声底但参数随血压漂移」这一格在脉象数据上尚未被直接观测到,目前由谐波失真随收缩压线性下降(PMID 35433650)间接推出,属合理推断而非实测。30 人 × 隔日 3 次的样本量与重复次数取自同类信度研究惯例,未针对脉象谐波做功效计算。

    待验证:脉象谐波残差在各阶的可吸收性分布(多少阶能被单参数吸收、多少阶须多参数补偿);参数解脊线的实际形状与宽度;「末位参数」在脉象模型里的具体名单与排序;判据三格在真实数据上的落入比例。

    相关阅读

    证据等级: C
    参考文献: PMID 10468357, PMID 19505944, PMID 22343490, PMID 26011432, PMID 33035441, PMID 34758345, PMID 35433650, PMID 41838759
    L1 网络状态:「节点表按『易学/中医侧』『物理侧』『生理/体验侧』三类域分派;物理侧仅列『相位 / 节律 —— 周期性耦合的关键』,生理侧八项(HRV、皮温/皮电、呼吸率/呼吸模式、EMG、EEG、胃肠蠕动、气感报告、体质分型)全为可穿戴单点横断面观测量」——该主题(谐波残差的参数可吸收性与结构缺维度判别)尚未进入 L1 对应网络,边表止于 E-050「六十甲子纳音 | 对应 | 60 个频率」,现代证据栏原文「尚无任何可测来源」;L1 对「残差 / 可辨识 / 反解 / 参数 / 血管 / 弹性 / 硬度 / 脉象 / 谐波 / 模型 / 血压 / 校准 / 补偿」十三个关键词全份文件零命中。


    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。

    Balancing the ledger does not mean it is correct: can residuals be fixed by tuning parameters, or does the model lack a dimension?

    Evidence level: C

    A residual being absorbed by a parameter does not mean that parameter is correct—it may simply be compensating for a dimension that does not yet exist in the model.

    Main Text

    I. Two Paths to the Same “Fit”

    When a pulse waveform fits poorly, leaving behind a residual segment, only two paths lie before us: either a parameter value is incorrect and needs adjustment, or the model itself lacks a dimension, making it impossible to fit regardless of tuning. The former is a minor repair; the latter requires major revision.

    These two paths lead to opposite conclusions. If the issue is merely lack of individualization of parameters, the resonance cavity hypothesis remains standing. However, if the missing element is a structural dimension, then ratios such as “what proportion is geometric and what proportion is elastic” remain entirely undefined until the structural term is added. The difficulty lies in the fact that both paths appear identical on the surface: “after adjustment, the curve fits better.” Distinguishing them requires a new set of criteria.

    II. Identifiability Has Two Layers

    In cardiovascular pharmacokinetic modeling, there is a clean case: a nonlinear hemodynamic feedback model always has parameters that cannot be estimated accurately. Structural identifiability analysis confirms the model itself is structurally unidentifiable; only after reparameterization (reducing the parameter table) does it become globally identifiable. Crucially, the two models before and after reparameterization are completely indistinguishable in their input-output behavior (PMID 22343490). This statement is significant: comparing curves alone can never distinguish between “the version with more parameters” and “the version with fewer parameters.”

    The Profile Likelihood approach splits these two layers more cleanly: structural unidentifiability manifests as functionally related parameter combinations—swapping one set of parameters yields the same output; practical unidentifiability arises from insufficient data quantity and quality (PMID 19505944). The former cannot be saved by adding samples, while the latter can. Therefore, “residuals that cannot be suppressed” must first be divided into two layers of inquiry, otherwise everything that follows is pure speculation.

    III. Practical Layer Has Hard Numbers

    Systematic testing of closed-loop lumped-parameter models provides a set of directly usable reference values: without noise, all parameters can be estimated with errors below one ten-thousandth; once 5% Gaussian noise is added, the errors of the five most sensitive parameters are limited to around 10%, while the four least sensitive parameters cannot be reliably estimated using waveform data, and the three least sensitive ones are “particularly hard to estimate”—the authors conclude prioritize measuring them rather than tuning them (PMID 34758345). The same study also found that using pressure waveforms as the cost function is several orders of magnitude better than using clinical indices such as stroke volume or pulse pressure.

    These three numbers provide a movable criterion: if residuals cannot be suppressed to the noise floor, first check where this parameter ranks in the sensitivity ordering. Those at the bottom being unsuppressible is normal, belonging to “not measured enough,” and does not constitute evidence that “the model is wrong.”

    IV. The Signature of Structural Error Is Wrong Direction

    Experiments on nonlinear elastic tubes provide a very clean signature. Using the same tube with two different physiological pulse amplitudes, linear methods show a significant increase in higher-order scattering, and as pulse amplitude rises, the three-point method yields physically untenable negative attenuation in lower orders (PMID 10468357). Negative attenuation coefficients have no physical explanation—that is an error in the model form, not an error in parameter values.

    This marks the dividing line between two types of errors. Parameter errors result in bounded residuals with generally random direction; structural errors cause residuals to cross into physically impossible quadrants. Therefore, the first question of the criterion should not be “how large is the residual,” but rather “has it run into impossible quadrants.”

    V. Conversely, It Is More Dangerous: The Better the Fit, the More Compensation Across Bars

    Harmonic distortion linearly decreases over 24 hours as systolic pressure rises (PMID 35433650), indicating amplitude dependence is an independent dimension; without it in the model, it emerges as “some parameter drifting with blood pressure.” Meanwhile, sympathetic activity independently regulates aortic stiffness under conditions unrelated to hemodynamics (PMID 33035441), effectively adding another free knob specifically to cancel out other parameters. More fundamentally: when arteriolar myogenic and metabolic regulation is limited, local resistance is determined by blood apparent viscosity rather than vessel diameter, and apparent viscosity itself varies with perfusion pressure and blood flow (PMID 26011432)—”resistance” is not constant physiologically.

    Therefore, “parameters being tuned correctly” must come with conditions: under what physiological state they are correct and for how long. If a geometric parameter drifts on an hourly scale with blood pressure, it is not being tuned correctly; it is working to compensate for the missing amplitude-dependence dimension. Multiple solutions under correlated inputs and parameter compensation further distort ratio estimates (PMID 41838759).

    VI. Three-Grid Criteria

    First grid, structural layer judgment first: after reparameterization, input-output behavior is indistinguishable, or residuals fall into physically impossible quadrants (negative attenuation, non-monotonicity) → judge structural dimension deficiency, requiring major overhaul. All ratio numbers are void; must first add structural terms such as pressure dependence, frequency-dependent viscosity before discussing allocation.

    Second grid, practical layer: single-parameter scanning can suppress the residual of that order to the repeat-measurement noise floor, and after suppression, parameter values remain stable across overnight repeats → judge parameters not individualized, requiring minor repair, hypothesis retained.

    Third grid is most prone to misjudgment: residuals suppressed to the noise floor, but parameter values drift significantly with instantaneous blood pressure or heart rate → judge successful absorption but wrong attribution. This grid must be listed separately because although it appears “fixed,” in reality it attributes the variance of missing structure to the name of a false parameter.

    VII. Measurement Method

    With ≥30 subjects, each undergoing repeat measurements three times overnight, estimate the noise floor and retest stability of parameter values for each order residual. For each order residual, sequentially perform three tasks: single-parameter scanning (can it be suppressed to the noise floor), multi-parameter joint inversion (does the solution drift along the ridge, is it unique), and correlation between residuals and instantaneous blood pressure/heart rate (direct reading of amplitude dependence). Sensitivity ranking is calculated first, with bottom-ranked parameters pre-classified as “not measured enough,” to avoid contaminating the judgment of “model is wrong.”

    One sentence conclusion: residuals are receipts, telling you the accounts are not balanced, but not where the money was spent. To judge minor or major repair, one must look at under what conditions parameters are correct, rather than how correct they are when they happen to be right.

    Evidence and Limitations

    Verified: Structural identifiability and practical identifiability must be handled separately in methodology, and models before and after reparameterization can be completely indistinguishable in input-output behavior (PMID 22343490, PMID 19505944); with 5% noise, the least sensitive parameters in cardiovascular lumped-parameter models indeed cannot be reliably estimated using waveform data (PMID 34758345); nonlinear elastic tubes yield physically untenable negative attenuation under high pulse amplitudes using linear methods (PMID 10468357); aortic stiffness can be independently regulated by sympathetic activity, and apparent viscosity is determined by perfusion pressure and blood flow (PMID 33035441, PMID 26011432). These are methodological evidence for “how to judge what to do,” with high strength.

    Partially Verified: The third grid—”residuals suppressed to noise floor but parameters drift with blood pressure”—has not been directly observed in pulse data; currently inferred indirectly from harmonic distortion linearly decreasing with systolic pressure (PMID 35433650), representing reasonable inference rather than direct measurement. The sample size of 30 subjects × 3 repeats overnight and the repetition count follow conventions from similar reliability studies, without power calculation specifically for pulse harmonic analysis.

    To Be Verified: The absorbability distribution of pulse harmonic residuals across orders (how many orders can be absorbed by single parameters, how many require multi-parameter compensation); the actual shape and width of parameter solution ridges; the specific list and ranking of “bottom-ranked parameters” in pulse models; the proportion of cases falling into each of the three grids under real data.

    Related Reading

    Evidence Level: C
    References: PMID 10468357, PMID 19505944, PMID 22343490, PMID 26011432, PMID 33035441, PMID 34758345, PMID 35433650, PMID 41838759
    L1 Network Status: “Node tables are assigned across three domains: ‘Yijing Studies / TCM side’, ‘Physics side’, and ‘Physiological/Experience side’; the Physics side lists only ‘Phase / Rhythm — key periodic coupling elements’, while the Physiological side includes eight items (HRV, skin temperature/skin conductance, respiration rate/respiratory pattern, EMG, EEG, gastrointestinal motility, qi-sensation reports, constitution typing) as wearable single-point cross-sectional observables.” This topic (parameter absorbability of harmonic residuals and structural dimension deficiency judgment) has not yet entered the corresponding L1 network. The edge table stops at E-050 “Sixty Jiazi Na Yin | Correspondence | 60 Frequencies,” with the original text of the modern evidence column stating “no measurable source yet available.” The L1 network has zero hits for all thirteen keywords: residual, identifiability, inversion, parameter, vessel, elasticity, hardness, pulse, harmonic, model, blood pressure, calibration, compensation.


    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.

  • 「一脏一谐波」的表该空腹查还是饭后查:重分配能不能把两个器官在谐波空间里换个位置

    Evidence level: C

    如果一顿饭就能把五个器官在谐波空间里的相对位置重排,那这张表就不是一张表,是一沓按时相分册的表——而过去所有研究都只拍了一张。

    正文

    一、那两条线,是拍的

    上一篇把 C1-C10 拆成逐器官裁决,留下一个必须先答的问题:器官血流占比在分钟到小时级就在变,那「一脏一谐波」是空腹表还是全天表。

    配的判据是:餐后与空腹两次的谐波剖面之差,若小于器官间差异的 30%,静态表成立;超过 60% 则彻底不成立。这两条线没有任何出处,是照抄信度研究里「三分之一等效界」的习惯写法。它可以先当占位符,不是结论。

    二、三条支路,同一顿饭,三种命运

    最直觉的做法是看肠系膜上动脉涨多少:口服 75 g 葡萄糖后,肠系膜上动脉平均血流升约 70%,门静脉升约 40%(PMID 40445944)。

    但信息在另外两条支路。腹腔干在 Fontan 病人身上餐后血管舒张直接消失,而肠系膜上动脉升幅与健康组无异(PMID 32857602);肝动脉的餐后缓冲反应在各种条件下都被保住,原文写「conserved under all conditions」(PMID 15086358)。

    三条支路、同一顿饭、三种反应。重分配不是整体平移的水位,是各支路各走各路。用「整体重分配幅度」一个数判表能不能用,方向上就错了——必须逐支路算改写量,再比它与邻支路的可分辨距离。

    三、最锋利的一刀:肝动脉是阴性对照

    那条「conserved」是一份礼物:它给了这个设计一条已知不该变的支路。

    判据因此可以改硬——不看绝对值,看模型对阴性对照报了什量。如果同一模型在肠系膜上动脉报出大改写、在肝动脉上也报出大改写,那改写量的来源不是分配,而是共模:波反射、僵硬度、心率这些作用于所有支路的因素被算进了「器官身份」。

    这不是假设。混合餐后 180 分钟内增强指数从 14% 降到 2%,脉搏波速度从 7.1 降到 6.7 m/s(PMID 24242825),而谐波剖面正由波反射与出口阻抗共同决定。输出确实变了,但从一根桡动脉上读不出是哪个输入变了——分配变了还是回波变了,在这个观测点上共线。

    四、漏掉的变量:上一顿是什么时候吃的

    原设计对照是「空腹 12 小时 vs 餐后 60 分钟」,里头藏着一个没锁的变量:距上一餐多久。

    24 名健康年轻人的实测给了答案:午餐后肠系膜上动脉血流在吃没吃早餐两种情况下都显著上升,但颈总动脉血流只在没吃早餐时才显著下降,吃了早餐就不降(PMID 30515951)。同一个人、同一顿午餐,上一餐的有无改掉了另一条支路的反应方向。

    所以「餐后状态」不是一个时点,是「距上一餐时长」的连续函数。这个变量没锁死之前,两点法测到的差值里混着上一餐的残留。

    五、60 分钟不是同一个生理时刻

    胃排空速率个体差异大,且随体重增加而加快(PMID 17087952)。同样标着「餐后 60 分钟」,在一个人身上是峰值,在另一个人身上已开始回落。

    更麻烦的是,健康人里有一部分会在进食后立即出现一次短暂的肠系膜血流跃升(PMID 38622189)。单点采样要么撞上要么错过——而撞上与错过,会被记成两个人「器官映射不同」。

    六、餐前流量已经把答案写进去了

    同一次研究还报:餐后肠道血流升幅的个体间变异很大,其中约 30% 可由餐前血流解释(PMID 38622189)。

    这对模型是致命的。若入口条件带餐前血流,而反应大小又部分由餐前血流决定,那模型输出的「稳定性」里有相当一部分是入口条件自带的,不是映射本身的稳定性。这制造伪稳定:换个时相测剖面也差不多,因为两次入口条件差不多——不是因为器官身份恒定。

    七、测法与四格判据

    不另招人,并入主标定:≥40 名健康受试者,同日三次全流程(空腹 ≥12 小时、标准餐后 20 分钟抓那记瞬升、餐后 60 分钟),另设「早餐后 3 小时再进午餐」对照臂。每次采五条供血动脉入口几何、桡动脉 ≥60 拍 FFT、肠系膜上动脉与腹腔干血流速度,同一 1-D 模型输出三时相逐器官剖面,以肝动脉为阴性对照。

    一:肠系膜改写量超其可分辨距离 60%、且肝动脉对照低于 20% ⇒ 重排成立,恒定映射不成立,只能作条件映射,未控进食时相的旧结论降级。二:两支路改写量相近且都大 ⇒ 改写来自共模不是分配,要撤的是「谐波剖面反映器官分配」这个中介假设,不是给表加时相标签。三:肠系膜 30%–60%、肝动脉低于 20% ⇒ 建空腹/餐后两套表,跨研究统一采血时相,「距上一餐时长」列为必报变量。四:三时相都低于 30% ⇒ 静态表成立,注明「空腹采」,30% 这条线可由实测反推标定。

    证据与局限

    已验证:餐后内脏血流大幅重分配——肠系膜上动脉升 70%(PMID 40445944)、腹腔干反应可被病理单独抹掉(PMID 32857602)、肝动脉缓冲在各条件下守恒(PMID 15086358)、混合餐显著降低增强指数与脉搏波速度(PMID 24242825)。这些是设计与判据的物理前提。

    部分验证:重分配可测但跨支路不同步(PMID 33547918 用 4D flow MRI 同测八条血管,餐后仅腹腔干以上主动脉、肠系膜上动脉、肠系膜上静脉、门静脉上升);上一餐改变下一餐反应方向(PMID 30515951);个体变异中 30% 由餐前血流解释(PMID 38622189)。

    待验证:① 重分配能否大到让两个器官在谐波空间互换位置——没有任何研究把餐后血流分配与脉搏谐波剖面同时测过;② 30%/60% 两条线无出处,须由四格实测反推标定;③ 本文引用证据全在血流层面,从血流份额到「第几阶谐波」这一跳仍缺可测中介;④ 阴性对照(肝动脉)能否从桡动脉单点信号里分离,尚未验证——若不能分离,第二、三格判据无法执行。

    L1 网络状态:该主题尚未进入 L1 对应网络。CORRESPONDENCE_NETWORK.md 生理侧节点表仅「心率变异性 / 皮温皮电 / 呼吸率 / 肌电 / EEG / 胃肠蠕动 / 气感报告 / 体质分型」八项,无「器官血流分配」「内脏血流份额」「餐后时相」任何节点;物理侧仅「相位 / 节律 — 周期性耦合的关键」一行。EDGE_TABLE_science-yixue.md 概念表中「五脏」26 篇最高 B 级、A 级数为 0,「六腑」4 篇 B 级;边表止于 E-050(纳音↔频率,现代证据栏原文「尚无任何可测来源」)。三域 grep「谐波 / 器官 / 血流 / 脏腑 / 时相 / 餐」仅命中五脏节点行与相位备注行,无任何一条边承载「器官↔谐波阶次」的时相依赖性。

    相关阅读

    证据等级: C
    参考文献: PMID 40445944, PMID 32857602, PMID 15086358, PMID 38622189, PMID 30515951, PMID 24242825, PMID 17087952, PMID 33547918


    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。

    Should the \”one organ, one harmonic\” table be checked fasting or after meals: can redistribution swap the positions of two organs in harmonic space

    Evidence level: C

    If a single meal could rearrange the relative positions of five organs within harmonic space, this table would not be a single chart but a stack of time-phased sections—yet all prior studies have captured only one snapshot.

    Body

    I. Those Two Lines Were Captured

    The previous analysis decomposed C1–C10 into organ-specific adjudications, leaving a prerequisite question unanswered: since organ blood flow proportion fluctuates from minute-to-hour scales, is the “one organ, one harmonic” model a fasting chart or an all-day chart?

    The adjudication criterion is as follows: the difference between postprandial and fasting harmonic profiles, if less than 30% of inter-organ variation, validates a static table; exceeding 60% renders it entirely invalid. These two lines have no provenance; they are copied from the habitual phrasing of “one-third equivalence threshold” in reliability studies. They may serve as placeholders, not conclusions.

    II. Three Branches, One Meal, Three Fates

    The most intuitive approach examines the increase in the superior mesenteric artery: after oral administration of 75 g glucose, superior mesenteric artery blood flow rises approximately 70%, and the portal vein rises approximately 40% (PMID 40445944).

    However, information exists on two additional branches. The celiac trunk’s postprandial vasodilation disappears entirely in Fontan patients, while the superior mesenteric artery increase matches the healthy group (PMID 32857602); the hepatic artery’s postprandial buffering response is conserved under all conditions (PMID 15086358).

    Three branches, one meal, three reactions. Redistribution is not a uniform water-level shift but each branch follows its own path. Using “overall redistribution magnitude” as a single metric to judge table applicability is directionally incorrect—one must calculate the rewrite amount for each branch individually and compare it with the resolvable distance of adjacent branches.

    III. The Sharpest Cut: Hepatic Artery as Negative Control

    That “conserved” finding is a gift: it provides this design with a known branch that should not change.

    The criterion can thus be hardened—ignoring absolute values, we examine what the model reports for the negative control. If the same model indicates large rewriting in both the superior mesenteric artery and the hepatic artery, the source of the rewrite is not distribution but common mode: factors such as wave reflection, stiffness, and heart rate acting on all branches have been incorporated into “organ identity.”

    This is not an assumption. Mixed postprandial enhancement index dropped from 14% to 2%, and pulse wave velocity decreased from 7.1 to 6.7 m/s within 180 minutes (PMID 24242825), while harmonic profiles are jointly determined by wave reflection and exit impedance. Output indeed changed, but from a single radial artery reading, one cannot discern which input changed—whether distribution or echo—since they are collinear at this observation point.

    IV. The Missing Variable: When Was the Last Meal?

    The original design contrasted “fasting 12 hours vs. postprandial 60 minutes,” concealing an unlocked variable: time since the last meal.

    Testing on 24 healthy young adults yielded the answer: superior mesenteric artery blood flow rose significantly whether breakfast was eaten or not, but common carotid artery flow dropped significantly only when breakfast was omitted; if breakfast was consumed, it did not drop (PMID 30515951). For the same person and same lunch, the presence or absence of the previous meal reversed the reaction direction of another branch.

    Thus “postprandial state” is not a time point but a continuous function of “time since last meal.” Until this variable is locked, the difference measured by two-point sampling mixes residual effects from the previous meal.

    V. 60 Minutes Is Not the Same Physiological Moment

    Gastric emptying rates vary greatly among individuals and accelerate with increased body weight (PMID 17087952). Even labeled as “60 minutes postprandial,” it represents the peak in one person but has already begun to decline in another.

    More problematically, some healthy individuals experience an immediate transient surge in superior mesenteric blood flow immediately after eating (PMID 38622189). Single-point sampling either hits or misses—yet hitting or missing is recorded as “different organ mapping” between two individuals.

    VI. Pre-meal Flow Already Wrote the Answer

    The same study also reported: individual variation in postprandial intestinal blood flow increase is large, with approximately 30% explainable by pre-meal blood flow (PMID 38622189).

    This is fatal for the model. If the entry condition includes pre-meal blood flow, and response magnitude is partially determined by pre-meal blood flow, then the “stability” in model output largely stems from entry conditions rather than mapping stability. This creates pseudo-stability: measuring profiles at different time points yields similar results because entry conditions are similar—not because organ identity is constant.

    VII. Measurement Method and Four-Grid Criteria

    Without recruiting additional subjects, we integrate into the main calibration: ≥40 healthy participants underwent three full cycles on the same day (fasting ≥12 hours, standard postprandial 20-minute capture of the transient surge, and postprandial 60 minutes), plus a control arm of “3 hours after breakfast before lunch.” Each session measured five arterial entry geometries, radial artery ≥60 FFT captures, superior mesenteric artery and celiac trunk blood velocities, with the same 1-D model outputting three time-phased organ profiles and using the hepatic artery as negative control.

    I: Superior mesenteric rewrite exceeds its resolvable distance by 60%, and hepatic artery control remains below 20% ⇒ rearrangement holds, constant mapping fails; only conditional mapping applies, and old conclusions uncontrolled for meal timing are downgraded. II: Both branches show similar large rewriting ⇒ the rewrite originates from common mode, not distribution; what must be withdrawn is the mediating assumption that “harmonic profiles reflect organ distribution,” not adding time-phased labels to the table. III: Superior mesenteric 30%–60%, hepatic artery below 20% ⇒ construct separate fasting/postprandial tables, unify blood sampling time across studies, and list “time since last meal” as a mandatory reported variable. IV: All three time phases below 30% ⇒ static table holds, noted as “fasting sample,” with the 30% threshold inferred from actual measurement calibration.

    Evidence and Limitations

    Verified: Postprandial visceral blood flow undergoes massive redistribution—superior mesenteric artery increases 70% (PMID 40445944), celiac trunk response can be erased by pathology alone (PMID 32857602), hepatic artery buffering is conserved under all conditions (PMID 15086358), and mixed meals significantly reduce enhancement index and pulse wave velocity (PMID 24242825). These are the physical premises of the design and criteria.

    Partially Verified: Redistribution is measurable but asynchronous across branches (PMID 33547918 used 4D flow MRI to simultaneously measure eight vessels, with only the aorta above the celiac trunk, superior mesenteric artery, superior mesenteric vein, and portal vein rising postprandially); the previous meal alters the reaction direction of the next meal (PMID 30515951); and 30% of individual variation is explained by pre-meal blood flow (PMID 38622189).

    To Be Verified: ① Can redistribution be large enough to swap the positions of two organs in harmonic space—no study has simultaneously measured postprandial blood flow distribution and pulse harmonic profiles; ② the 30%/60% lines lack provenance and must be calibrated via four-grid actual measurements; ③ all cited evidence in this article pertains only to the blood flow level, lacking a measurable intermediary from blood share to “which harmonic order”; ④ whether the negative control (hepatic artery) can be separated from single-point radial artery signals remains unverified—if separation fails, criteria II and III cannot be executed.

    L1 Network Status: This topic has not yet entered the L1 corresponding network. The physiological side node table in CORRESPONDENCE_NETWORK.md includes only eight items: heart rate variability, skin temperature/skin conductance, breathing rate, electromyography, EEG, gastrointestinal motility, qi sensation report, and constitution typing—no nodes for “organ blood flow distribution,” “visceral blood share,” or “postprandial time phase”; the physical side contains only one row: “phase/rhythm—key of periodic coupling.” The concept table in EDGE_TABLE_science-yixue.md shows “five zang” with 26 highest B-level entries and zero A-level, while “six fu” has four B-level entries; the edge table stops at E-050 (Na Yin ↔ Frequency, with original text stating “no measurable source yet”). Grep across three domains for “harmonic / organ / blood flow / zang-fu / time phase / meal” hits only the five zang node row and the phase note row, with no edge carrying “organ ↔ harmonic order” time-phase dependency.

    Related Reading

    Evidence Level: C
    References: PMID 40445944, PMID 32857602, PMID 15086358, PMID 38622189, PMID 30515951, PMID 24242825, PMID 17087952, PMID 33547918


    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.

  • 哪个器官的出口阻力根本反解不出来:五床逐床体检,以及「敏感」并不等于「可辨识」

    Evidence level: C

    五个床的阻力里,可能有两个在数学上根本没有唯一解——这不是数据不够,是加多少样本都救不回来的那一种。

    一、判据把三件事混成了一件事

    504 的整个逐器官裁决压在一个前提上:五个床的出口阻力都能从静息桡动脉波形里唯一反解出来。#1177 提出用三个指标体检——参数扫描看谱变化幅度(敏感度)、多初值反解看是否收敛到同一解(唯一性)、隔日重测看 ICC。三项并列、任一不过就判「不可辨识」。

    这个并列是错的。三件事在数学上是三个层级,混在一起会产生两种误判。

    二、先例:四元件模型里的 L 就是反解不出来

    有直接证据。三元件与四元件 Windkessel 用子空间法辨识时,三元件没有显著系统误差,而四元件对动脉惰性 L 的估计出现显著系统误差(P=0.011),原文结论是「四元件模型的应用因此存疑」,并指出噪声水平升高时离群值增加「几乎全部来自 L 的估计误差」(PMID 20172779)。

    注意这里只有一根管子、两个待解参数。一个在单管模型里就结构上辨识不住的参数,放到五个并联床上不会变好。

    三、最像五床的那个拓扑,恰恰被判非全局可辨识

    另一条更贴。用微分代数方法对心血管模型做先验全局可辨识性分析时,系统动脉循环的 Windkessel 与 T-Tube、心肌收缩的几种模型全部通过,唯一被判非全局可辨识的就是 T-Tube 模型(PMID 26726299)。

    T-Tube 是什么?一个入口、两个出口。这正是五床结构的最小类比。而五床就是 T-Tube 再加三支。若两支已经不可辨识,五支的可辨识性不能靠「多做几个人」补上。

    四、五个床其实只有四个自由度

    还有一层更基本的约束被跳过了。桡动脉给出的是总流量与总压力,五个床的阻力以并联方式合成一个总阻力。总阻力一旦被观测定住,五个床只剩四个可独立变动的比值——第五个由其余四个决定。

    也就是说「五床全部可辨识」在结构上就达不到,除非观测里带进了各支路的分流信息(波形形状、相位关系)。而这些恰恰是信噪比最低的部分。原判据里「若五床全部可辨识则 #504 无保留成立」这一格,可能从设计之初就是空集。

    五、要改判的一条:敏感度低不是不可辨识

    这是本文最要紧的一处。

    参数扫描谱变化 <5% 被判为「不可辨识」,逻辑不成立。谱变化小说明这个量对输出几乎没影响,它可能完全可辨识——只不过辨识得再准也没用,因为输出看不见它。后果不是「无法比对」,而是「该床的命中不算证据」:模型无论怎么调都会「命中」,这种命中没有信息量。

    两者豁免的理由不同,处理也不同。真不可辨识要把它从比对里剔除;敏感度低要保留比对,只在统计时标注「该床不提供支持性证据」。原判据把后者也判成豁免,等于用一条错误的理由删掉一个该留的床。

    六、信度数字不能跨条件比较

    ICC <0.6 这条线也需限定采集条件。肾动脉用 3T 螺旋相位速度成像测 RI,两次检查差值 −0.00±0.04(PMID 25648103),看起来极稳。但另一组 18 例移植肾用常规双功超声测叶间动脉 RI,结论只是「观察者间变异大于观察者内变异,但均在 BSI 采纳的重复性界限内」(PMID 8330135)。两者不是一个量级的方法,数字不能混用。

    更麻烦的是静默失败。肾动脉正常波形的早期收缩峰出现率本身受血管顺应性调制,顺应性高的年轻人常看不到这个峰(PMID 7644627)。若反解算法依赖这个特征,它会在相当一部分人身上失效——但不报错,只给出一个看似正常的解。

    反过来,肝动脉 RI 单次测量标准差 0.08、内在变异占总方差 76%(PMID 8756922),说明同一床在不同条件下的信度可以差一个量级。而边界阻力的影响在冠脉仿真里仅次于几何,排在粘度与病变长度之前(PMID 26803339)——这一项不小。

    七、测法:结构层先做,人后做

    顺序要倒过来。第一步是纯计算的:把五床拓扑写进代数可辨识性判定,看哪几个床在结构上就有解不唯一——这一步不需要受试者,且结论不受样本量影响。第二步才上人:≥60 名受试者,对通过结构判定的床做多初值反解(≥100 组随机初值)与隔日重测,每床另做一次生理范围参数扫描记录谱变化幅度。

    判据分四格:结构不可辨识 ⇒ 直接豁免,不进统计;结构可辨识但多解比例 >20% 或 ICC <0.6 ⇒ 实践不可辨识,豁免并注明「换观测点可救」(如相位对比 MRI 逐分支流量,PMID 22771032 的非线性方程组加敏感性分析是这条路线的现成模板);可辨识但敏感度 <5% ⇒ 保留比对、命中不计为支持证据;四项全过 ⇒ #504 该床的裁决无保留成立。

    证据与局限

    已验证:四元件模型的 L 存在低可辨识与系统性估计误差(PMID 20172779);T-Tube 拓扑在全局可辨识性分析中被判非全局可辨识(PMID 26726299);边界阻力对血流仿真的影响量级仅次于几何(PMID 26803339)。

    部分验证:肝动脉 RI 的个体内变异占 76%(PMID 8756922)与肾动脉 RI 在严格控制下差值近零(PMID 25648103)并存,说明「信度」高度依赖采集方法而非器官本身;肾动脉早期收缩峰的出现率随顺应性变化(PMID 7644627)提示部分床存在静默失败风险。

    待验证:五床并联是否真的只余四个自由度,需按具体拓扑做一次代数判定确认;「敏感度低 ≠ 不可辨识」这一改判在真实五床模型上的后果量级尚未计算;结构判定为可辨识、而实践中仍大面积多解的床是否存在,尚无数据。

    相关阅读

    L1 网络状态:CORRESPONDENCE_NETWORK.md 物理侧仅列「相位 / 节律 —— 周期性耦合的关键」这一指标名,不带角度性、不带误差带、不带度量空间;生理侧八项(HRV、皮温 / 皮电、呼吸率 / 呼吸模式、EMG、EEG、胃肠蠕动、气感报告、体质分型)全为可穿戴单点观测项,唯一与血流相关的是「皮温 / 皮电 | 可穿戴 | 情绪/血流窗口」,无「器官出口阻力 / 血管床 / 参数反解 / 可辨识性」任何节点或边;EDGE_TABLE_science-yixue.md 边表止于 E-050(「六十甲子纳音 | 对应 | 60 个频率」,层级 L1、状态 open),现代证据栏原文「尚无任何可测来源」,全文无 L1/L2/L3 与实际状态判定。该主题尚未进入 L1 对应网络。

    证据等级: C
    参考文献: PMID 20172779, PMID 26726299, PMID 22771032, PMID 26803339, PMID 8756922, PMID 7644627, PMID 25648103, PMID 8330135


    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。

    Which organ’s exit resistance cannot be fundamentally solved: five beds examined one by one, and \”sensitive\” does not equal \”identifiable\

    Evidence level: C

    Among the resistance of five beds, two may have no unique mathematical solution at all—not due to insufficient data, but because adding more samples cannot recover them.

    I. The Criterion Mixes Three Distinct Matters into One

    504’s entire organ-by-organ adjudication rests on a single premise: the outlet resistance of all five beds can be uniquely inversely solved from the resting radial artery waveform. #1177 proposes examining three indicators—parameter scanning to observe spectral variation amplitude (sensitivity), multi-initial-value inversion to check convergence to the same solution (uniqueness), and retesting the next day to assess ICC. These three are treated as parallel; failure of any one leads to a judgment of “not identifiable.”

    This parallelism is incorrect. In mathematics, these represent three distinct levels; conflating them produces two types of misjudgment.

    II. Precedent: In the Four-Component Model, L Cannot Be Inversely Solved

    Direct evidence exists. When three-component and four-component Windkessel models are identified via subspace methods, the three-component model shows no significant systematic error, whereas the four-component model exhibits significant systematic error in estimating arterial inertance L (P=0.011); the original conclusion states “the application of the four-component model is therefore questionable,” noting that as noise levels rise, outliers increase “almost entirely due to estimation error of L” (PMID 20172779).

    Note here there is only one tube and two parameters to solve. A parameter that cannot be structurally identified even in a single-tube model will not become identifiable when placed into five parallel beds.

    III. The Topology Most Similar to Five Beds Is Itself Judged Non-Global Identifiable

    Another point is even more relevant. When performing a priori global identifiability analysis on cardiovascular models using differential-algebraic methods, the arterial circulation Windkessel and T-Tube models, along with several myocardial contraction models, all pass; the only one judged non-globally identifiable is the T-Tube model (PMID 26726299).

    What is the T-Tube? A single inlet and two outlets. This precisely matches the minimal structural analogy of five beds. Five beds correspond to the T-Tube plus three additional branches. If two branches are already non-identifiable, the identifiability of five cannot be compensated by “adding more people.”

    IV. Five Beds Actually Possess Only Four Degrees of Freedom

    A more fundamental constraint has been overlooked. The radial artery provides total flow and total pressure; the resistances of five beds combine in parallel to form a single total resistance. Once this total resistance is fixed by observation, only four ratios among the five beds can vary independently—the fifth is determined by the other four.

    Thus, “all five beds identifiable” is structurally impossible unless the observations include branching flow information (waveform shape, phase relationships). Yet these are precisely the parts with the lowest signal-to-noise ratio. The criterion cell stating “if all five beds are identifiable, then #504’s judgment holds unreserved” may have been an empty set from the design stage.

    V. One Criterion to Revise: Low Sensitivity Does Not Mean Non-Identifiability

    This is the most critical point of this article.

    When spectral variation <5% via parameter scanning, it is judged as “not identifiable,” but the logic fails. Small spectral variation indicates this quantity has almost no effect on output; it may be fully identifiable—yet identifying it accurately is useless because the output cannot reflect it. The consequence is not “cannot be compared,” but rather “this bed’s hit does not count as evidence”: regardless of how the model is tuned, it will “hit,” but such hits carry no information.

    The reasons for exemption differ between the two cases, and so do the treatments. Truly non-identifiable beds must be excluded from comparison; low-sensitivity beds must remain in comparison, with statistical annotation stating “this bed does not provide supporting evidence.” The original criterion incorrectly treats the latter as exempt, effectively deleting a necessary bed under a flawed rationale.

    VI. Reliability Numbers Cannot Be Compared Across Conditions

    The threshold ICC <0.6 also requires limitation of acquisition conditions. Renal artery RI measured via 3T spiral phase-contrast imaging shows a difference of −0.00±0.04 between two examinations (PMID 25648103), appearing extremely stable. However, another study of 18 transplanted kidneys using conventional dual-frequency ultrasound reports inter-observer variation exceeding intra-observer variation, yet both within BSI’s accepted repeatability limits (PMID 8330135). These two methods differ by orders of magnitude; their numbers cannot be mixed.

    More problematic is silent failure. The early systolic peak in normal renal artery waveforms is modulated by vascular compliance; young individuals with high compliance often lack this peak (PMID 7644627). If the inverse algorithm depends on this feature, it will fail for a substantial portion of subjects—not by error, but by yielding a seemingly normal solution.

    Conversely, hepatic artery RI shows a single-measurement standard deviation of 0.08 and intrinsic variation accounting for 76% of total variance (PMID 8756922), indicating reliability across conditions can differ by an order of magnitude. In coronary artery simulations, boundary resistance impact ranks second only to geometry, ahead of viscosity and lesion length (PMID 26803339)—a significant factor.

    VII. Measurement Order: Structure First, Humans Later

    The sequence must be reversed. Step one is purely computational: encode the five-bed topology into algebraic identifiability criteria to determine which beds lack unique solutions structurally—this step requires no subjects and its conclusion is unaffected by sample size. Step two then involves humans: ≥60 subjects undergo multi-initial-value inversion (≥100 random initial values) and next-day retesting for structurally identifiable beds, plus a physiological-range parameter scan per bed to record spectral variation amplitude.

    The criteria divide into four cells:
    – Structurally non-identifiable ⇒ direct exemption, no statistical entry;
    – Structurally identifiable but multi-solution ratio >20% or ICC <0.6 ⇒ practically non-identifiable, exempt with note “can be rescued by changing observation point” (e.g., phase-contrast MRI for branch-by-branch flow, PMID 22771032; nonlinear equation system with sensitivity analysis serves as the ready template for this route);
    – Identifiable but sensitivity <5% ⇒ retain comparison, hits do not count as supporting evidence;
    – All four pass ⇒ #504’s judgment for this bed holds unreserved.

    Evidence and Limitations

    Verified: L in the four-component model exhibits low identifiability and systematic estimation error (PMID 20172779); the T-Tube topology is judged non-globally identifiable in global identifiability analysis (PMID 26726299); boundary resistance impact on blood flow simulation ranks second only to geometry (PMID 26803339).

    Partially Verified: Individual variation in hepatic artery RI accounts for 76% (PMID 8756922) while renal artery RI difference under strict control approaches zero (PMID 25648103), indicating “reliability” heavily depends on acquisition method rather than organ itself; the appearance rate of early systolic peaks in renal artery waveforms varies with compliance (PMID 7644627), suggesting some beds carry silent failure risk.

    To Be Verified: Whether five parallel beds truly retain only four degrees of freedom requires algebraic confirmation via specific topology; the magnitude of consequence from revising “low sensitivity ≠ non-identifiability” on real five-bed models remains uncomputed; whether beds exist that are structurally identifiable yet practically multi-solution in large scale awaits data.

    Further Reading

    L1 Network Status: CORRESPONDENCE_NETWORK.md lists only the indicator name “phase/rhythm — key of periodic coupling,” without angularity, error bands, or metric space; the physiological side includes eight items (HRV, skin temperature/skin conductance, respiration rate/respiratory pattern, EMG, EEG, gastrointestinal motility, qi-sensation report, constitution typing), all single-point wearable observations; the only blood-flow-related item is “skin temperature/skin conductance | wearable | emotion/blood-flow window,” with no nodes or edges for “organ outlet resistance / vascular bed / parameter inversion / identifiability”; EDGE_TABLE_science-yixue.md stops at E-050 (“Sixty Jiazi Yin-Yang | Correspondence | 60 Frequencies”), level L1, status open; the original evidence column states “no measurable source yet exists,** and the full text contains no L1/L2/L3 or actual status determination. This topic has not yet entered the L1 correspondence network.

    Evidence Level: C
    References: PMID 20172779, PMID 26726299, PMID 22771032, PMID 26803339, PMID 8756922, PMID 7644627, PMID 25648103, PMID 8330135


    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.

  • 「知道在被测」到底值几个度:遮蔽与知情下静息相位角的差值量级,以及它相对等效界的位置

    Evidence level: C

    两态比较的准入条件压在一个没人量过的数字上:受试者知道自己在被记录呼吸,胸腹相位角会偏几度。这个数不测出来,1.67° 这条线就只是猜的。

    一、判据写得像规矩,底下的数是空的

    前面几篇把静息态与练功态的相位角比较,建在一条操作规矩上:若「知情 − 遮蔽」的静息 φ 差大于等效界三分之一(5° 的 1/3,即 1.67°),未遮蔽的基线就不可用。规矩清楚、可执行,但它压在一个从未被测过的量上,三分之一这个比例也没有出处。判据可以先拍,拍了必须补测。

    二、三条证据都说 φ 会动,但没有一条给度数

    74 名健康人做三条件记录:伪装成「设备校准,请等五分钟」、明确告知正在记录呼吸、口嘴加流速计。知晓被记录使吸气与呼气时间同时延长,叹息与呼气末停顿几近消失(PMID 9032510)。82 人随机分入知情与不知情组,知情组呼吸率平均低 2.13 次/分,且效应不随计数时长改变(PMID 28426897)。18 名呼吸模式障碍者接受单次实时相位角反馈,肋腹相位角从 −2.89° 走到 −1.39°(PMID 34073590)。

    三条都真实,但结局变量是秒和次/分,不是度。φ 是角度,不是周期的派生量——呼吸变慢不等于胸腹时移变大。这个数只能实测,不能换算。

    三、遮蔽拿掉的是「知晓」,不是「实验室」

    受试者可以不知道呼吸在被记录,但体位、操作员、胸带、实验室的时间都在。这类场所效应在血压上有硬数字:31 名 70 岁以上住院患者出院后做 24 小时动态血压,收缩压从 133.5 降到 126.2 mmHg,焦虑量表同步从 7.5 降到 5.0,且焦虑变化能预测收缩压变化(PMID 28834872)。这已是「尽量少打扰」的方案,仍有此量级落差。若呼吸上存在同构效应,「遮蔽 vs 知情」测到的就不是纯粹的知晓,而是「知晓 + 场所」的合量。

    四、同一偏倚在两种设计里指向相反结论

    若资深者的 φ 优势本就有一部分来自习惯性轻度接管,遮蔽把这条路径也切掉,两态差被压小,同样指向趋同——只是这次是遮蔽造成的。不遮蔽把差做大,遮蔽把差做小,同一偏倚方向相反。要判断该信哪个,先得知道这个差有多大、符号在个体间是否一致。

    五、这个差必须和「换算法能差多少」比

    呼吸信号的天然变异本就不小:5 名健康人同天两次、隔周两次的肺音谱,同天变异 32.8%(吸气)、隔周 36.9%,麦克风位置稍移可把局部变异推到 48%–78%(PMID 8306049)。φ 还多一层算法差:五种相位估计算法在同一批模拟信号上,傅里叶相位法算整段单一值零误差、算时变相位误差可观,悖动法与李萨如法平均误差约 4%(PMID 19963785)。两者同量级,就分不清是心理效应还是估计方法。

    六、1.67° 其实卡在信号本身的大小上

    OEP 静息时肋腹相位角健康组 −1.9°、呼吸模式障碍组 −2.7°,组间差 0.8°(PMID 38655165);单次反馈让 −2.89° 走到 −1.39°,差 1.5°(PMID 34073590);ALS 患者静息时各分区 θ 显著高于配对健康人(PMID 29969705)。可查的 φ 效应量都在 1–2°。1.67° 不是宽松阈值,它恰好等于一个临床组间差的量级。若测量偏倚与待测效应同量级,未遮蔽的两态比较就无法把效应从偏倚里捞出来。

    七、怎么测

    可与主标定共用受试者。40 名以上有丹田呼吸经验者,隔 2–7 天重复 4 次,每次同体位、同带位、同操作员,按随机顺序做两段各 10 分钟静息记录:一段完全遮蔽(受试者以为在做心率检查、呼吸带藏在衣服下、不做呼吸指导),一段完全知情(明确告知在评估呼吸模式、要求「像平时一样自然」),中间留 5 分钟洗脱。主结局为差值绝对值与符号;副结局为呼吸率、腹式占比、叹息次数;并用两种相位算法各算一遍。

    八、四格判据

    一:中位绝对值 >1.67° → 三分之一这条线被证实为合理甚至偏松,任何未遮蔽的静息基线不可用于两态比较,既往同类研究基线一律降级。

    二:落在 0.5°–1.67° → 偏倚存在但可当固定偏倚在分析中校正,不必整体重做。

    三:<0.5°(小于等效界十分之一)→ 遮蔽对 φ 不必要,这条判据应撤除,只保留对主观气感报告的口头指导控制。

    四:绝对值大但符号在个体间随机 → 它测的不是系统偏倚而是个体差异,遮蔽须改为按再投入倾向分层,而非统一遮蔽。

    证据与局限

    已验证:知晓呼吸被记录会改变呼吸模式与频率,有随机分组实验支撑(PMID 28426897);实时相位角反馈能在一次干预内改变 φ(PMID 34073590);场所与焦虑能改变生理读数(PMID 28834872)。

    部分验证:φ 各分区效应量在 1–2° 量级,来自运动态与静息态混合样本,静息态绝对值范围尚无健康人常模。

    待验证:「知情 − 遮蔽」的静息 φ 差值本身至今没有任何直接测量。三分之一这个比例、以及遮蔽是否必要,全部悬在这一步上。

    相关阅读

    L1 网络状态:该主题尚未进入 L1 对应网络。物理侧节点表仅有「相位 / 节律 — 周期性耦合的关键」一行,不带角度性、不含量纲与测量学维度;生理侧「呼吸率 / 呼吸模式 — 与节律耦合最直接」为横断面观测项;边表止于 E-050,其现代证据栏原文为「尚无任何可测来源」。

    证据等级: C
    参考文献: PMID 9032510, PMID 28426897, PMID 34073590, PMID 28834872, PMID 8306049, PMID 19963785, PMID 38655165, PMID 29969705


    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。

    How Many Degrees Does \”Known to Be Tested\” Amount to: Magnitude of the Difference in Resting Phase Angle Under Occlusion and Knowledge, and Its Position Relative to the Equivale…

    Evidence level: C

    The admissibility criterion for two-state comparison rests on a number never directly measured: when subjects know they are being recorded, their thoraco-abdominal phase angle shifts by a certain number of degrees. Without measuring this value, the 1.67° threshold remains merely a guess.

    I. The Criterion Is Written Like a Rule, But the Numbers Below Are Empty

    Previous articles compared resting-state and practice-state phase angles based on an operational rule: if the difference in resting φ between “informed” and “blinded” conditions exceeds one-third of the equivalence boundary (5° / 3 = 1.67°), then the unblinded baseline cannot be used. The rule is clear and executable, yet it hinges on a quantity never directly measured, nor does the proportion of one-third have any cited origin. The criterion can be set tentatively, but once set, it must be empirically supplemented.

    II. Three Lines of Evidence Confirm φ Shifts, But None Provide Degrees

    74 healthy participants underwent three-condition recording: disguised (“device calibration, please wait five minutes”), explicitly informed (breathing is being recorded), and mouthpiece with flow sensor. Knowledge of recording simultaneously prolonged both inspiration and expiration times, nearly eliminating sighs and end-expiratory pauses (PMID 9032510). In a randomized trial of 82 participants, the informed group had an average breathing rate 2.13 breaths/min lower than the unblinded group, with effects independent of counting duration (PMID 28426897). Eighteen individuals with breathing pattern disorders received single-session real-time phase-angle feedback, shifting rib-abdomen phase angle from −2.89° to −1.39° (PMID 34073590).

    All three findings are real, yet the outcome variables are seconds and breaths per minute, not degrees. φ is an angle, not a derived quantity of the cycle—slower breathing does not mean larger thoraco-abdominal time shifts. This value must be measured directly; it cannot be converted from rate data.

    III. Blinding Removes “Knowledge,” Not the “Lab”

    Subjects may not know their breathing is being recorded, yet posture, operator presence, chest straps, and laboratory timing remain constant. Such setting effects yield hard numbers in blood pressure: 31 hospitalized patients over 70 years old showed systolic BP dropping from 133.5 to 126.2 mmHg and anxiety scales from 7.5 to 5.0 over 24 hours post-discharge, with anxiety change predicting BP change (PMID 28834872). This is already a “minimize disturbance” protocol, yet such magnitude differences persist. If breathing exhibits analogous effects, the “blinded vs. informed” measurement captures not pure knowledge but the combined effect of “knowledge + setting.”

    IV. The Same Bias Points to Opposite Conclusions in Two Designs

    If experts’ φ advantage partly stems from habitual mild autotuning, blinding cuts off this pathway, compressing the two-state difference and pointing toward convergence—but now due to blinding. Unblinding amplifies the difference; blinding reduces it, creating opposite bias directions for the same underlying tendency. To judge which to trust, one must first know the magnitude of this difference and whether its sign is consistent across individuals.

    V. This Difference Must Be Compared Against “Conversion Method Error”

    Breathing signals inherently vary widely: lung sound spectra from 5 healthy individuals measured twice on the same day and biweekly showed intra-day variation of 32.8% (inspiration) and inter-week variation of 36.9%; slight microphone repositioning could push local variation to 48%–78% (PMID 8306049). φ adds another layer of algorithmic error: five phase-estimation algorithms applied to the same simulated signals yielded zero error for Fourier-based whole-segment estimation but significant error for time-varying phase, while paradoxical and Lissajous methods averaged ~4% error (PMID 19963785). Both sources are of comparable magnitude, making it impossible to distinguish psychological effects from estimation method errors.

    VI. 1.67° Actually Hinges on the Size of the Signal Itself

    In OEP resting-state rib-abdomen phase angle, healthy groups averaged −1.9° and breathing-pattern-disordered groups −2.7°, a group difference of 0.8° (PMID 38655165); single-session feedback shifted −2.89° to −1.39°, a difference of 1.5° (PMID 34073590); ALS patients showed significantly higher θ values across partitions than matched healthy controls (PMID 29969705). Measurable φ effect sizes all fall within 1–2°. The 1.67° threshold is not a loose margin—it exactly matches the magnitude of clinical group differences. If measurement bias and target effect are of the same order, unblinded two-state comparison cannot extract the true effect from the bias.

    VII. How to Measure

    This can share subjects with the main calibration. Over 40 individuals with Dantian breathing experience repeat measurements every 2–7 days, four times each, maintaining identical posture, strap placement, and operator, randomized into two 10-minute resting recordings: one fully blinded (subjects believe they are undergoing heart rate testing, breathing strap hidden under clothing, no breathing guidance) and one fully informed (explicitly told breathing pattern is being assessed, instructed to breathe “as naturally as usual”), with a 5-minute washout period in between. Primary outcome is the absolute difference and sign; secondary outcomes include breathing rate, abdominal breathing proportion, and sigh count; both phase algorithms are applied.

    VIII. Four-Grid Criterion

    I: Median absolute difference >1.67° → The one-third threshold is confirmed reasonable or even loose; any unblinded resting baseline cannot be used for two-state comparison, and all previous similar studies’ baselines must be downgraded.

    II: Falls within 0.5°–1.67° → Bias exists but can be treated as fixed bias and corrected in analysis; full re-measurement is unnecessary.

    III: <0.5° (less than one-tenth of the equivalence boundary) → Blinding is unnecessary for φ; this criterion should be removed, retaining only verbal guidance control on subjective qi sensation reports.

    IV: Large absolute value but random sign across individuals → It measures individual differences, not systematic bias; blinding must shift to stratification by re-engagement tendency rather than uniform blinding.

    Evidence and Limitations

    Verified: Knowledge of breathing recording alters breathing pattern and rate, supported by randomized trials (PMID 28426897); real-time phase-angle feedback changes φ within a single intervention (PMID 34073590); setting and anxiety alter physiological readings (PMID 28834872).

    Partially Verified: φ partition effect sizes range 1–2°, derived from mixed dynamic and resting samples; no healthy population norm exists yet for absolute resting values.

    Pending Verification: The resting φ difference between “informed” and “blinded” conditions has never been directly measured. The one-third proportion and necessity of blinding hinge entirely on this step.

    Related Reading

    L1 Network Status: This topic has not yet entered the corresponding L1 network. The physical-side node table contains only “Key of Phase/Rhythm — Periodic Coupling,” lacking angular specificity, dimensional units, and measurement dimensions; the physiological side lists “Breathing Rate/Breathing Pattern — Most Directly Coupled with Rhythm” as a cross-sectional observation item; the edge table stops at E-050, with its modern evidence section originally stating “no measurable source yet.”

    Evidence Level: C
    References: PMID 9032510, PMID 28426897, PMID 34073590, PMID 28834872, PMID 8306049, PMID 19963785, PMID 38655165, PMID 29969705


    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.

  • 呼吸肌从来没有真正「停练」过:把「撤训练」和「撤使用」分开后,低剂量维持组能不能把衰减砍掉一半

    Evidence level: C

    所有停练研究都默认「不练了」等于「刺激归零」,但呼吸肌做不到归零——它必须一直工作。这个被忽略的差额,可能才是「十年功力」里最容易被误读的那一半。

    正文

    第一层:一个从未被检验的假设。 停练研究的整套设计建立在一句话上:撤掉训练,等于撤掉刺激。四肢肌大致成立,你可以真的躺着不动。呼吸肌不成立——哪怕什么都不练,它每天仍要收缩两万次左右维持肺泡通气。于是「停练」在呼吸侧从来不是零剂量,而是一个未知的低剂量。既往研究把这个未知量当成了零。

    第二层:三组证据互相打架,说明「停练」不是一个变量。 完全反向组:老年女性 4 周吸气肌训练后停练 4 周,最大吸气压回落 23±8 cmH₂O、自主神经指标同步退回基线,等于全部归还(PMID 33864832)。保住组:5 周训练后停练 5 周,吸气压仍高于训练前(PMID 36754374);停练 9 周只小幅下降并进入平台、仍高于基线(PMID 12569211);晚发型 Pompe 病 12 周训练后停 3 个月,「大体耐久」(PMID 26381077)。减量组:训练频率砍掉三分之二仍能维持到停练后 18 周(PMID 12569211);把频率降到每两周甚至每四周一次、强度与量不变,12 周后各角度力量无显著下降,而彻底停练组掉了约 55%(PMID 1471008);每周一次即可保住肩袖力量增益,每周零次才出现下降(PMID 11078906)。

    同一个现象,三种结局。合理的解释不是谁做错了,而是「停练」这个词把两件事压成了一个:撤训练(特定负荷刺激消失)与撤使用(日常通气负荷水平发生变化)。这两件事方向可以相反——停练后体能下降、体重上升、焦虑增加,日常呼吸做功反而可能升高。若如此,「停练期」的净刺激未必下降,甚至可能上升。

    第三层:撤使用是真的,而且很快。 控制性机械通气把膈肌去负荷后,肌内基因表达谱在数小时内成批改变(PMID 16126937)。更关键的一刀在这里:即使在压力支持通气下保留部分吸气努力,只要负荷水平足够低,膈肌功能与结构照样退化(PMID 22425820)。「在用」不等于「在训」。 这正好是呼吸肌停练研究的软肋——它把「有自主呼吸」当成了「有训练刺激」。

    第四层:反向的一刀,测量本身会动。 最大吸气压有明确的学习效应:吸气肌热身的效应量 0.40(95%CI 0.17–0.63),显著高于单次测试的 0.20(0.05–0.35)与重复测试的 0.14(0.07–0.22)(PMID 35504306)。COPD 患者的测试-重测信度研究同样记录了学习效应(PMID 8339633);预先做亚极量呼吸肌活动可削弱这一效应(PMID 11223726)。含义很硬:维持组「没掉」,可能有相当一部分不是肌肉保住了,而是第二次测得更准、更会用力的结果。三臂的差异若小于这个量级,任何解读都不成立。

    第五层:减量组的成绩单,反而把「老本」的边界打糊了。 每四周一次就能保住力量,说明维持所需的刺激门槛低到接近于零。低到与「日常呼吸习惯」无法区分时,「十年功力」和「十年呼吸习惯」在数据上就是同一个东西。这不是修辞,是设计问题:如果维持剂量与日常负荷同量级,那么任何「停练后仍高于常人」的观察,都无法排除「因为他本来就每天这样呼吸」。

    第六层:怎么测。 三臂随机对照:完全停练 / 每周 2 次、每次 5 分钟的极低剂量维持 / 继续常规训练,各 ≥30 人,训练期时长与峰值严格匹配;另设一个从未训练的对照组作基准,否则「高于基线」可能只是「高于他自己的低起点」。停练 6/12/24 周测吸气压、口腔闭合压、最大吸气流速与呼吸肌耐力,统一按绝对变化量比较,并在每次测试前执行同一热身流程、取多次最大值。缺了热身统一,学习效应就会伪装成组间差异。

    第七层:四种结果怎么读。 低剂量组降幅比完全停练组小一半以上 ⇒ 衰减里有一大块其实是日常呼吸习惯改变,不是功力流失,「老本」应改称「维持剂量效应」,既往停练研究的剂量归零假设集体失效。三组无差异 ⇒ 呼吸肌确实存在撤刺激即衰减的通路,老本只能来自训练史本身。完全停练组反而掉得更少 ⇒ 先查维持组的额外注意负荷与测压学习效应,别急着下结论。三组都在测量误差带内(约 0.14–0.40 效应量)⇒ 这个实验在现有工具分辨率下做不出来,得先解决测量问题。

    证据与局限

    已验证:减量训练可维持力量增益(每周一次至每四周一次均有实测支撑);完全停练的衰减幅度显著大于减量训练;机械通气去负荷会快速改变膈肌基因表达并致萎缩;最大吸气压存在可量化的学习效应与热身增益。

    部分验证:呼吸肌停练后的保留时长,各项研究报告从「4 周完全归还」到「3 个月大体耐久」跨度极大,样本量普遍偏小(多为 8–24 人),且训练强度、停练定义、测量流程均不统一,无法合并。

    待验证:「撤训练」与「撤使用」的分离从未被直接实验过;极低剂量维持的剂量-反应曲线没有标定点;停练期日常通气负荷是否真的变化,没有任何研究测过。

    本主题的证据整体薄弱:上述结论多为小样本单一研究,且「未真正归零」这一前提本身只是推理而非实测,本文不构成对任何练功时长的机制性证明。

    相关阅读

    证据等级: C
    参考文献: PMID 1471008, PMID 11078906, PMID 12569211, PMID 36754374, PMID 33864832, PMID 26381077, PMID 16126937, PMID 22425820, PMID 35504306, PMID 8339633, PMID 11223726

    L1 网络状态:「该主题尚未进入 L1 对应网络」——CORRESPONDENCE_NETWORK.md 节点表生理侧只列「呼吸率 / 呼吸模式」为可穿戴观测项(备注「与节律耦合最直接」)与「气感报告」为结构化自评,二者均未与易学侧任何节点连成边;节点表内不存在「训练剂量」「停练 / 撤使用」「维持负荷」这类纵向干预维度节点。边表 3.1–3.4 只登记五行↔五脏/五色/五音/五方/五季/五志、八卦↔五行、天干地支↔五行,无任何以呼吸肌功率、训练剂量或停练为源或目标的边。EDGE_TABLE_science-yixue.md 仅有概念计数(呼吸 189 篇 / 气功 48 篇 / 丹田 51 篇),无层级与状态判定。故本文封顶 C。


    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。

    Respiratory muscles never truly rest: Can low-dose maintenance cut the decline in half by separating detraining from disuse?

    Evidence level: C

    All detraining studies default to the assumption that “no training” equals “stimulus zero,” but breathing muscles cannot reach zero—they must always work. This overlooked gap may be the most easily misread half of the “decade-long cultivation” narrative.

    Main Text

    Layer 1: An Unverified Hypothesis. The entire design of detraining studies rests on one premise: removing training equals removing stimulus. This holds roughly true for limb muscles, where one can simply lie still. It does not hold for breathing muscles—even with no formal training, they contract roughly 20,000 times daily to maintain alveolar ventilation. Thus, “detraining” on the respiratory side is never zero dose; it is an unknown low dose. Prior research treated this unknown quantity as zero.

    Layer 2: Three Lines of Evidence Conflict, Showing “Detraining” Is Not a Single Variable.
    * Full Cessation Group: After 4 weeks of inspiratory muscle training followed by 4 weeks of cessation, maximal inspiratory pressure (MIP) dropped 23±8 cmH₂O and autonomic indices returned to baseline simultaneously, indicating full regression (PMID 33864832).
    * Retention Group: After 5 weeks of training followed by 5 weeks of cessation, MIP remained higher than pre-training levels (PMID 36754374); after 9 weeks of cessation, only a slight decline occurred before plateauing, still above baseline (PMID 12569211); in late-onset Pompe disease, after 12 weeks of training followed by 3 months of cessation, there was “substantial durability” (PMID 26381077).
    * Dose-Reduction Group: Reducing training frequency by two-thirds still maintained gains for 18 weeks post-cessation (PMID 12569211); reducing frequency to every two or four weeks while keeping intensity and volume constant resulted in no significant decline in strength from any angle after 12 weeks, whereas the full cessation group lost approximately 55% (PMID 1471008); weekly sessions were sufficient to preserve rotator cuff strength gains, while zero sessions resulted in decline (PMID 11078906).

    The same phenomenon yields three different outcomes. A reasonable explanation is not that one group made an error, but that the term “detraining” compresses two distinct events into one: removing training (disappearance of specific load stimulus) and removing use (change in daily ventilatory load levels). These two events can move in opposite directions—post-cessation physical performance may decline, weight may rise, and anxiety may increase, while daily respiratory work could actually rise. If so, the net stimulus during the “cessation period” may not decrease; it could even increase.

    Layer 3: Removing Use Is Real, and It Happens Quickly. Controlled mechanical unloading of the diaphragm causes batch changes in intramuscular gene expression within hours (PMID 16126937). A more critical finding here is that even with partial inspiratory effort retained under pressure support ventilation, if the load level is sufficiently low, diaphragm function and structure still degenerate (PMID 22425820). “Being used” does not equal “being trained.” This is precisely the Achilles’ heel of breathing muscle detraining research—it equates “having spontaneous breathing” with “having training stimulus.”

    Layer 4: The Reverse Cut, Measurement Itself Moves. MIP exhibits a clear learning effect: the effect size for inspiratory muscle warm-up is 0.40 (95%CI 0.17–0.63), significantly higher than single-test at 0.20 (0.05–0.35) and repeated-test at 0.14 (0.07–0.22) (PMID 35504306). Studies on test-retest reliability in COPD patients also recorded learning effects (PMID 8339633); pre-exercise submaximal breathing muscle activity can attenuate this effect (PMID 11223726). The implication is stark: the maintenance group “did not drop” may have a substantial portion of that stability due to more accurate or forceful second testing rather than preserved muscle. If the inter-arm difference is smaller than this magnitude, no interpretation holds.

    Layer 5: The Dose-Reduction Group’s Scorecard Blurs the Boundary of “Old Stock.” The ability to maintain strength with sessions every four weeks indicates that the stimulus threshold required for maintenance is near zero. When this is low enough to be indistinguishable from “daily breathing habits,” “decade-long cultivation” and “decade-long breathing habit” become the same thing in the data. This is not rhetoric; it is a design flaw: if the maintenance dose is commensurate with daily load, any observation of “remaining above average post-detraining” cannot rule out “he simply breathes like this daily.”

    Layer 6: How to Measure. Three-arm randomized control trial: Full cessation / Very low-dose maintenance (twice weekly, 5 minutes each) / Continued conventional training; each arm ≥30 participants, with training duration and peak intensity strictly matched. Include a control group of never-trained individuals as a baseline; otherwise, “above baseline” may simply mean “above their own low starting point.” Measure MIP, oral closure pressure, maximal inspiratory flow rate, and breathing muscle endurance at 6/12/24 weeks post-cessation. Compare all changes absolutely, and execute the same warm-up protocol before each test, taking multiple maximums. Without standardized warming, learning effects will masquerade as inter-group differences.

    Layer 7: How to Read Four Possible Outcomes.
    * Low-dose group shows >50% less decline than full cessation ⇒ A large chunk of the decay is actually due to changes in daily breathing habits, not loss of “old stock”; “old stock” should be renamed “maintenance dose effect,” and the zero-dose assumption of prior detraining studies collectively fails.
    * Three groups show no difference ⇒ There is indeed a pathway where removing stimulus leads to decay; “old stock” can only come from training history itself.
    * Full cessation group drops less than expected ⇒ First investigate additional attentional loads in the maintenance group and measurement learning effects before drawing conclusions.
    * All three groups fall within the measurement error band (~0.14–0.40 effect size) ⇒ This experiment cannot be performed with current tool resolution; measurement issues must be resolved first.

    Evidence and Limitations

    Verified: Dose-reduced training can maintain strength gains (supported by empirical data for weekly to every-four-week sessions); full cessation shows significantly greater decay than dose-reduced training; mechanical unloading rapidly alters diaphragm gene expression and causes atrophy; MIP exhibits a quantifiable learning effect and warm-up gain.

    Partially Verified: The duration of retention post-breathing muscle cessation varies wildly across reports, ranging from “4 weeks full regression” to “3 months substantial durability.” Sample sizes are generally small (mostly 8–24 participants), and training intensity, definitions of cessation, and measurement protocols are not unified, preventing meta-analysis.

    To Be Verified: The separation of “removing training” from “removing use” has never been directly tested experimentally; there are no anchor points for the dose-response curve of very low-dose maintenance; whether daily ventilatory load truly changes during cessation has never been measured in any study.

    Overall Evidence for This Topic Is Weak: The above conclusions rely largely on small-sample single studies, and the premise of “never truly reaching zero” itself is inferential rather than empirically measured. This article does not constitute mechanistic proof for any specific cultivation duration.

    Related Reading

    Evidence Level: C
    References: PMID 1471008, PMID 11078906, PMID 12569211, PMID 36754374, PMID 33864832, PMID 26381077, PMID 16126937, PMID 22425820, PMID 35504306, PMID 8339633, PMID 11223726

    L1 Network Status: “This topic has not yet entered the L1 Correspondence Network”—The CORRESPONDENCE_NETWORK.md node table lists physiological side only “Respiration Rate / Breathing Pattern” as wearable observation items (noted for “most direct coupling with rhythm”) and “Qi Sensation Reports” as structured self-assessments; neither connects to any node on the Yijing side. No longitudinal intervention dimension nodes such as “Training Dose,” “Cessation/Detachment of Use,” or “Maintenance Load” exist within the node table. Edge Tables 3.1–3.4 only register Five Phases ↔ Five Organs/Five Colors/Five Tones/Five Directions/Five Seasons/Five Emotions, Eight Trigrams ↔ Five Phases, Heavenly Stems and Earthly Branches ↔ Five Phases; there are no edges sourced or targeted by breathing muscle power, training dose, or cessation. EDGE_TABLE_science-yixue.md contains only concept counts (Breathing: 189 papers / Qigong: 48 papers / Dantian: 51 papers), with no hierarchy or status determination. Therefore, this article is capped at Level C.


    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.