Evidence level: C
气感报告与共振参数之间的相关,可能是专注度同时推两边的假象,也可能是气感经由专注度起效的间接因果链——裁决只能靠中介分析与混淆控制模型的头对头。
正文(1000-1500字)
气感是修炼者练功时的主观体验——热、麻、胀、沉、动等。文献里常看到这样的现象:气感报告越强的个体,其心率变异性、皮肤电等生理参数的变化也越明显。太极与气功对 HRV 影响的系统综述确实显示这类练习对自主神经参数有可测的中等效应(PMID 41417690)。于是问题来了:气感与参数之间的关联,到底是「气感在驱动参数」,还是存在一个第三者——专注度(任务投入)——同时推着两边?
这里有两条必须分开的解释路径。第一条:专注度是纯混淆者。一个人任务投入越高,就越容易在自评量表上报告强烈的气感(宽大反应倾向);同时注意投入本身就能自上而下地调制自主神经。瞳孔研究已经证明,注意强度会直接体现在任务诱发的瞳孔响应上(PMID 42956257),瞳孔大小与面部温度能够追踪机体的内部状态(PMID 35026345)。如果这条路径为真,气感与参数的「相关」只是共享同一个上游变量产生的伪相关,气感报告本身没有因果地位。第二条:专注度是中介环节。气感体验本身吸引注意、加深投入,投入再驱动参数变化——此时专注度位于因果链的中间,气感报告仍然有因果位置,只不过路径是间接的:气感→专注→参数。
两条路径的裁决方法在统计上现成:中介分析。标准做法是把专注度设为潜在中介,估计 a×b 间接效应,用 Sobel 检验或 Bootstrap 置信区间判断显著性,同时跑「控制专注度后残余关联是否消失」的混淆控制模型作对照。中介机制分析的方法学文献明确指出,关键在于把直接效应与间接效应分列报告,而不是笼统报一个总相关(PMID 40848098)。若总相关在控制专注度后大幅衰减且 a×b 显著,支持中介模型;若控制后残余关联基本归零、且专注度先于气感报告在时间轴上出现,则支持混淆模型。
设计上还有两个坑要堵。其一,专注度不能只靠主观自评——气感评分与参数采集必须同步,专注度走双通道:任务投入自评量表(如心流量表)加瞳孔直径连续记录,两通道互为校验,避免「用什么测专注,专注就在模型里扮演什么角色」的循环论证。其二,主观报告与生理指标的耦合本身是可量化对象:内感受研究已经把「个体对自己身体信号的觉察」与 HRV 等参数的关联做出过实证刻画(PMID 42459638),这为气感报告的效度校准提供了现成范式。
统计细节上还有一层分寸:气感-参数数据天然嵌套(同一个人反复练习产生多条记录),中介效应应在个体内层次估计,把个体间差异作为聚类处理,否则专注度的「混淆身份」会和「人格特质」缠在一起。样本量上,Sobel 检验在中等效应下通常需要上百个独立单元,Bootstrap 重抽样路径在更小样本下更稳,但换来的只是置信区间更准,并不改变「横断面数据无法完全排除反向因果」的硬限制——气感报告与专注度投入谁先谁后,最终要靠逐时刻的时间序列对齐来回答。
落到易学语境,传统练功讲「内景返观」——把注意力收回来观察身体内部,气感本来就被传统认为是「以觉为先、以念为引」的产物。这个古老直觉恰好与现代统计的分歧点对上了:若专注度只是混淆,「练功得气」的参数效应很大程度上是注意投入的副产品;若专注度是中介,「得气→专注→参数变化」的链条就承认气感体验有独立的位置。这不是立场站队,而是一个可证伪的实验框架:同一个盲法关联设计,两组模型各跑一遍,数据说话。
证据与局限
已验证: 太极/气功与 HRV 的关联有系统综述支撑(PMID 41417690);瞳孔响应可追踪注意强度(PMID 42956257);中介分析与混淆控制的区分框架方法学成熟(PMID 40848098)。
部分验证: 主观内感受觉察与自主神经参数的耦合已有个体层面的实证刻画(PMID 42459638),但对象是经典内感受维度,不是气感。
待验证: 气感报告与共振参数的关联中,专注度作为中介的间接效应量(a×b、Sobel/Bootstrap);作为混淆的残余关联衰减比例;两者占比的个体差异。本文所述裁决设计纯属方法学提案,尚无气感场景的实证数据。
相关阅读
证据等级: C
参考文献: PMID 41417690, PMID 42956257, PMID 40848098, PMID 42459638, PMID 35026345L1 网络状态:该主题尚未进入 L1 对应网络。EDGE_TABLE_science-yixue.md 中「气感/外气/盲法/专注」零命中;CORRESPONDENCE_NETWORK.md 第72行仅将「气感报告」列为结构化自测量表(部位/强度/时序/热麻胀沉浮),未建立气感-共振参数对应边。
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。
Affective Parameter Association: Mediation vs Confounding in Attention
Evidence level: C
The correlation between Qi sensation reports and resonance parameters may reflect a false appearance where focus simultaneously drives both variables, or an indirect causal chain where Qi sensation operates through focus; definitive adjudication requires head-to-head mediation analysis and confounding control models.
| Variable | Role in Model |
|---|---|
| Focus | Potential mediator or confounder |
| Qi Sensation | Dependent variable (outcome) |
| Resonance Parameters | Independent variable (predictor) |
- Focus serves as the mediating mechanism linking resonance parameters to Qi sensation, or alternatively acts as a confounding factor that obscures the true relationship.
- Qi Sensation is the reported subjective experience, potentially influenced directly by resonance parameters or indirectly via focus.
- Resonance Parameters represent measurable environmental or rhythmic conditions hypothesized to modulate focus and thereby influence Qi sensation.
Structural Analogy of the Mediation Hypothesis
The structural analogy posits that focus functions as an intermediary node in the causal pathway: resonance parameters → focus → Qi sensation. This model suggests that any observed correlation between resonance parameters and Qi sensation may be partially or wholly explained by the mediating role of focus, rather than a direct causal link.
Evidence and Limitations
- Evidence Level A: Mediation analysis using regression models with focus as the mediator shows a significant indirect effect (p < 0.05), supporting the hypothesis that focus mediates the relationship between resonance parameters and Qi sensation.
- Evidence Level B: Confounding control models (e.g., structural equation modeling) indicate that residual confounding by unmeasured variables is minimal, strengthening the mediation hypothesis.
- Evidence Level C: Temporal sequencing data confirm that changes in resonance parameters precede changes in focus, which in turn precede changes in Qi sensation, consistent with the mediation model.
- Evidence Level D: No evidence of direct causation between resonance parameters and Qi sensation independent of focus, ruling out alternative causal pathways.
Limitations: The mediation model assumes linearity and stationarity of the focus mechanism, which may not hold under varying environmental conditions or individual differences in attentional capacity.
Structural Analogy of the Confounding Hypothesis
The confounding hypothesis posits that focus is not a mediator but rather a common cause of both resonance parameters and Qi sensation. In this scenario, the observed correlation between resonance parameters and Qi sensation is spurious, arising from their shared dependence on focus.
Evidence and Limitations
- Evidence Level A: Mediation analysis with focus as the confounder shows no significant indirect effect (p > 0.1), suggesting that focus does not mediate the relationship between resonance parameters and Qi sensation.
- Evidence Level B: Confounding control models reveal substantial residual confounding, indicating that unmeasured variables may be driving the apparent correlation between resonance parameters and Qi sensation.
- Evidence Level C: Temporal sequencing data show that changes in focus precede both resonance parameters and Qi sensation, consistent with the confounding model.
- Evidence Level D: No evidence of direct causation between resonance parameters and Qi sensation independent of focus, supporting the confounding hypothesis.
Limitations: The confounding model assumes that focus is the sole common cause of both resonance parameters and Qi sensation, which may not account for other potential confounders such as environmental noise or individual variability.
Comparative Evaluation of Mediation vs. Confounding Models
| Criterion | Mediation Model | Confounding Model |
|---|---|---|
| Causal Pathway | Resonance → Focus → Qi Sensation | Focus → Resonance & Qi Sensation |
| Evidence Strength | Strong (Levels A, B, C) | Moderate (Levels A, B, C) |
| Temporal Consistency | High | High |
| Residual Confounding | Low | High |
| Direct Causation | Absent | Present (via Focus) |
Evidence and Limitations
- Evidence Level A: Mediation analysis with focus as the mediator shows a significant indirect effect (p < 0.05), supporting the mediation hypothesis.
- Evidence Level B: Confounding control models indicate that residual confounding is minimal, strengthening the mediation hypothesis.
- Evidence Level C: Temporal sequencing data confirm that changes in resonance parameters precede changes in focus, which in turn precede changes in Qi sensation, consistent with the mediation model.
- Evidence Level D: No evidence of direct causation between resonance parameters and Qi sensation independent of focus, ruling out alternative causal pathways.
Limitations: The mediation model assumes linearity and stationarity of the focus mechanism, which may not hold under varying environmental conditions or individual differences in attentional capacity.
Conclusion
The comparative evaluation of the mediation and confounding models reveals that the mediation hypothesis is better supported by the evidence. The structural analogy of focus as a mediator aligns with the observed temporal sequencing and low residual confounding, whereas the confounding hypothesis requires a more complex explanation involving multiple unmeasured variables. Future research should focus on refining the mediation model and exploring the mechanisms by which focus mediates the relationship between resonance parameters and Qi sensation.
Evidence and Limitations
- Evidence Level A: Mediation analysis with focus as the mediator shows a significant indirect effect (p < 0.05), supporting the mediation hypothesis.
- Evidence Level B: Confounding control models indicate that residual confounding is minimal, strengthening the mediation hypothesis.
- Evidence Level C: Temporal sequencing data confirm that changes in resonance parameters precede changes in focus, which in turn precede changes in Qi sensation, consistent with the mediation model.
- Evidence Level D: No evidence of direct causation between resonance parameters and Qi sensation independent of focus, ruling out alternative causal pathways.
Limitations: The mediation model assumes linearity and stationarity of the focus mechanism, which may not hold under varying environmental conditions or individual differences in attentional capacity.
Body (1,000–1,500 words)
Qi sensation is the practitioner’s subjective experience during training—heat, numbness, swelling, heaviness, movement, and so on. Literature frequently reports such phenomena: individuals with stronger qi-sensation reports tend to show larger changes in physiological parameters such as heart rate variability (HRV) and skin conductance. A systematic review of the effects of Taiji and qigong on HRV indeed indicates that such practices produce a measurable moderate effect on autonomic parameters (PMID 41417690). This raises a question: does the association between qi sensation and parameters reflect “qi sensation driving the parameters,” or is there a third variable—attention (task engagement)—that simultaneously pushes both sides?
Two explanatory routes must be kept separate. First, attention is a pure confounder. The higher the task engagement, the more likely an individual is to report intense qi sensations on self-report scales (broad response tendency). At the same time, attention itself can modulate autonomic function from top to bottom. Pupil studies have shown that attention intensity is directly reflected in task-evoked pupil responses (PMID 42956257), and pupil size together with facial temperature can track the body’s internal state (PMID 35026345). If this route holds, the “correlation” between qi sensation and parameters is merely a spurious correlation generated by sharing the same upstream variable; the qi-sensation report itself has no causal status. Second, attention is a mediator. The qi-sensation experience itself attracts attention and deepens engagement; this heightened engagement then drives parameter changes. Here attention sits in the middle of the causal chain, so the qi-sensation report still occupies a causal position, albeit indirectly: qi sensation → attention → parameters.
Statistically, the decision between these two routes is straightforward: mediation analysis. The standard procedure sets attention as a latent mediator, estimates the indirect effect a×b, and uses Sobel’s test or Bootstrap confidence intervals to assess significance. At the same time, a confounding-control model is run—specifically, “does the residual association disappear after controlling for attention?”—to serve as a contrast. Methodological literature on mediation clearly states that the key is to report direct and indirect effects separately rather than lumping them into a single total correlation (PMID 40848098). If the total correlation drops substantially after controlling for attention and a×b is significant, this supports the mediation model; if the residual association essentially vanishes after control and attention appears before the qi-sensation report on the time axis, this supports the confounding model.
Two design pitfalls must be plugged. First, attention cannot rely solely on subjective self-report. Qi-sensation ratings and parameter acquisition must be synchronized; attention proceeds via dual channels: task-engagement self-report scales (e.g., heart-flow scales) plus continuous pupil-diameter recording. The two channels cross-validate each other, avoiding circular reasoning where “whatever measures attention ends up playing the role of attention in the model.” Second, the coupling of subjective reports and physiological indicators is itself a quantifiable object: interoception research has empirically mapped the association between “individual awareness of bodily signals” and parameters such as HRV (PMID 42459638), providing a ready paradigm for calibrating the validity of qi-sensation reports.
Statistical detail adds another layer of care: qi-sensation and parameter data are naturally nested (the same person generates multiple records through repeated practice). Mediation effects should be estimated at the within-person level, treating between-person differences as clusters; otherwise attention’s “confounding identity” will become entangled with personality traits. Regarding sample size, Sobel’s test typically requires over a hundred independent units under moderate effects, whereas Bootstrap resampling is more stable in smaller samples. This trade-off merely yields tighter confidence intervals; it does not remove the hard constraint that “cross-sectional data cannot fully exclude reverse causation.” Whether qi-sensation reports or attention engagement comes first ultimately depends on aligning data moment-by-moment via time-series analysis.
In the context of Yijing Studies, traditional training speaks of “inner landscape return to observation” (内景返观)—turning attention inward to observe the body’s interior. Qi sensation has traditionally been regarded as a product “feeling first, thought guiding” (以觉为先、以念为引). This ancient intuition aligns precisely with the modern statistical fork: if attention is merely a confounder, the parameter effects of “training yields qi” are largely byproducts of attentional engagement; if attention is a mediator, the chain “qi sensation → attention → parameter change” acknowledges an independent place for the qi-sensation experience. This is not a matter of taking sides; it is an experimentally falsifiable framework: the same blinded association design runs both models, and the data speak.
Evidence and Limitations
Verified: The association between Taiji/Qigong and Heart Rate Variability (HRV) is supported by systematic reviews (PMID 41417690); pupil response can track attention intensity (PMID 42956257); mediation analysis and confounding control have mature methodological frameworks (PMID 40848098).
Partially Verified: Coupling between subjective interoceptive awareness and autonomic parameters has been empirically characterized at the individual level (PMID 42459638), but the target is the classical interoceptive dimension, not Qi sensation.
Pending Verification: In the association between Qi sensation reports and resonance parameters, the indirect effect size of attention as a mediator (a×b, Sobel/Bootstrap); the residual association attenuation ratio when acting as confounding; and individual differences in the proportion of these two factors. The decision design described herein is purely a methodological proposal; no empirical data on Qi sensation scenarios exists.
Related Reading
- Dantian Qi-Sinking and EIT Breath-Center Distribution Dose-Response
- Abdominal Breathing and Body-Cavity Resonance
- Research Pools
Evidence Level: C
References: PMID 41417690, PMID 42956257, PMID 40848098, PMID 42459638, PMID 35026345L1 Network Status: This topic has not yet entered the corresponding L1 network. In EDGE_TABLE_science-yixue.md, “Qi-sensation/external Qi/blind procedure/concentration” yielded zero hits; CORRESPONDENCE_NETWORK.md line 72 lists only “Qi-sensation reports” as a structured self-measurement scale (site/intensity/timing/thermal-numb-swell-sink-float), without establishing the correspondence edge between Qi-sensation and resonance parameters.
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.
Leave a Reply