The Meridian-Program Hypothesis: Rethinking the Acupuncture System as a Bioelectrical Control and Information Network

1. Introduction

The meridian theory of traditional Chinese medicine describes a body-wide regulatory network linking body surface sites, internal organs and functional states. In modern biomedical language, however, this theory has often been interpreted as implying discrete anatomical conduits. Because conventional dissection and imaging have not identified a separate meridian tube system, the meridian concept has remained scientifically controversial. A more productive approach is to ask whether the classical meridian construct may correspond to a distributed physiological control system whose existence is expressed through measurable relations, state changes and regulatory effects rather than through a single visible organ.

This reframing is consistent with the history of information and control sciences. Shannon separated the engineering problem of signal transmission from the problem of meaning [1]. MacKay later emphasized that the meaning of a message is the change it produces in the receiver [2]. Wiener placed communication and control within a single scientific framework, showing that purposeful biological and mechanical behavior can be understood through feedback [3]. For acupuncture research, this distinction is crucial: the central question is not only whether a signal can travel through tissue, but how a living system interprets that signal and converts it into regulated physiological output.

We therefore propose the Meridian-Program Hypothesis. The term 'program' is used as a disciplined systems metaphor, not as a claim that the body literally runs computer code. It highlights three properties that are central to acupuncture: inputs are parameterized, signals are context-dependent, and outputs are produced through layered interpretation and feedback. This article synthesizes existing evidence and translates the hypothesis into testable predictions.

2. The Meridian-Program Hypothesis

The Meridian-Program Hypothesis conceptualizes the acupuncture system as a dynamic bioelectrical and mechanochemical control network. It has five interacting layers: hardware, code, compiler, execution and state correction. The hardware layer includes connective tissue planes, fascia, interstitial microenvironments and specific sensory afferent pathways. The code layer consists of multimodal signals generated by needling or related stimulation, including electrical impedance shifts, mechanical deformation, chemical mediator release and possibly ultraweak photon emission. The compiler layer refers to central nervous system interpretation, especially somatosensory, brainstem, autonomic and limbic processing. The execution layer includes downstream neuroimmune, endocrine, vascular and molecular pathways. The state-correction layer represents acupuncture as a controllable intervention that can modify input parameters and exploit feedback to move a dysregulated system toward a more stable state.

This framework differs from a simple pathway model. A pathway model asks whether stimulation at a point activates one anatomical route. A program model asks how multiple substrates cooperate to encode a signal, route it, interpret it and execute a state-dependent response. It also avoids overclaiming: not every traditional meridian statement must map one-to-one onto a single anatomical structure. Instead, classical meridian descriptions may be treated as historically accumulated maps of body-surface access points to regulatory functions.

2.1. Hardware Layer: Fascia, Interstitium and Neural Access Routes

Several lines of evidence support the plausibility of a hardware layer. Electrical impedance studies suggest that some acupuncture points and meridian-related regions have distinctive electrical properties, although methodological quality varies and findings should be interpreted cautiously [4,5,6]. Connective tissue research further indicates that mechanical stimulation during acupuncture can engage fibroblasts, extracellular matrix remodeling and tissue-level mechanotransduction [7,8]. An anatomical multicenter study reported relationships between meridian trajectories and extracellular matrix structures in human cadavers [9], while recent reviews emphasize fibroblast-mediated extracellular matrix organization as a candidate substrate for acupoint and meridian function [10].

At the neural level, electroacupuncture studies have identified somatotopic and intensity-dependent autonomic effects. Liu and colleagues showed that different body regions and stimulation intensities can drive distinct NPY-expressing sympathetic pathways [12], and later work demonstrated a neuroanatomical basis for electroacupuncture to drive the vagal-adrenal axis [13]. More recent work on gastric function suggests that deep tissue stimulation at ST36 recruits specific peripheral and central circuits in mice and humans [14]. Together, these findings suggest that the hardware layer is not a single tissue type but a composite access architecture involving fascia, interstitium and neural routes.

2.2. Code Layer: Multimodal Biological Signals

The code layer proposes that acupuncture inputs are encoded as multimodal biological signals. Electrical properties are one candidate channel. Systematic and experimental studies have reported lower impedance or altered electrical characteristics at some acupuncture-related sites [4,5,6], but heterogeneity in instrumentation, skin preparation and blinding means that stronger standardization is needed. Mechanical coding is another channel: needle rotation and tissue deformation can produce connective-tissue winding, fibroblast responses and downstream biochemical changes [7,8,11].

Chemical and molecular signals may also contribute. Reviews of acupuncture mechanisms describe local and systemic changes in neurotransmitters, neuropeptides, inflammatory mediators, ion channels, phosphorylation, methylation and histone modification [15,16]. An interdisciplinary review of meridian and acupoint characteristics summarizes electrical, thermal, optical, chemical and biological observations and argues that future work should move from macroscopic structure-seeking toward microstructural and functional integration [17]. In the present hypothesis, these signals are not treated as isolated mechanisms. They are candidate alphabets of a physiological code whose meaning depends on spatial location, stimulation parameters and body state.

2.3. Compiler Layer: Central Interpretation of Peripheral Inputs

If acupuncture signals are codes, the central nervous system is the principal interpreter. Functional neuroimaging has shown that stimulation at different acupoints can modulate partially distinct brain networks. Whole-brain functional connectivity analysis has been used to investigate acupoint specificity [18], and newer fMRI studies report sex-related differences in brain responses to LR3 stimulation [19]. A multimodal meta-analysis in healthy participants identified robust neural signatures of acupuncture modulation in core sensory and pain-processing networks [20].

The compiler metaphor is useful because the same peripheral input can have different effects depending on the receiver state. In biological terms, this means that pain, inflammation, autonomic tone, sex, disease duration, expectation and prior exposure can change the interpretation of an acupuncture signal. Therefore, lack of a uniform response across all subjects should not be treated only as noise. It may be an essential feature of a state-dependent control system.

2.4. Execution Layer: Molecular and Neuroimmune Interfaces

The execution layer consists of the downstream pathways that convert interpreted signals into physiological effects. Persistent pain research has implicated peripheral, spinal and supraspinal mechanisms of acupuncture and electroacupuncture [15]. Molecular reviews describe regulation of neuroimmune, neuroendocrine and vascular pathways, including HPA-axis signaling, cytokine modulation, neurotransmitter release and anti-inflammatory mechanisms [16,21,22]. From the viewpoint of the Meridian-Program Hypothesis, these pathways are analogous to biological execution interfaces: they do not explain the entire program, but they execute specific modules of the system.

This framing helps reconcile specificity and generality. Some acupuncture effects may be specific because they recruit particular sensory afferents, autonomic circuits or organ-related pathways. Other effects may be general because multiple inputs converge on shared immune, endocrine or pain-modulatory modules. A systems model can accommodate both types without forcing all evidence into either a highly specific meridian map or a nonspecific placebo/stress-response model.

2.5. State-Correction Layer: Acupuncture as Parameterized Feedback Intervention

The state-correction layer is the most speculative part of the model, but it is also the part with the clearest experimental consequences. In this layer, disease is treated as a dysregulated control state rather than as a single local lesion. Acupuncture is a parameterized input: point selection, depth, angle, manipulation, frequency, intensity, duration and treatment schedule all matter. A 'bug' in this context means a fault state in sensing, interpretation, execution or feedback, not a literal software error.

This layer aligns with contemporary moves toward quantified acupuncture manipulation and data-driven precision integrative medicine [23,24,25,26]. It also resonates with Chinese discussions of bidirectional regulation and meridian differentiation, which emphasize that the same intervention may normalize opposite pathological states depending on context [27,28,29]. A network meta-analysis of acupuncture for insulin resistance, for example, suggests that modality and dose may influence clinical outcomes [30]. Such findings do not prove the hypothesis, but they support the idea that stimulation parameters carry biologically meaningful information.

3. Evidence Integration

The value of the Meridian-Program Hypothesis lies in integration rather than replacement. Fascia-based models highlight physical continuity and mechanotransduction; neural models identify afferent and autonomic circuits; molecular models explain downstream effector pathways; neuroimaging models describe central representation; and AI-based precision medicine approaches offer tools for individualized prediction. Each is incomplete alone. A layered control model clarifies how these views can be mutually supportive.

Importantly, the evidence base remains uneven. Electrical point studies are suggestive but methodologically heterogeneous [4]. Cadaveric and fascia studies are anatomically informative but do not by themselves prove functional meridians [8,9,10]. High-quality neural circuit studies show strong causal mechanisms for specific electroacupuncture effects [12,13,14], but they do not establish that the full traditional meridian map is biologically fixed. Neuroimaging studies support network modulation but are often correlational and sensitive to analytical choices [18,19,20,23]. These limitations are reasons to refine the hypothesis, not to overstate it.

Figure 1. Five-layer architecture of the Meridian-Program Hypothesis.

Figure 1. Five-layer architecture of the Meridian-Program Hypothesis.

The framework treats acupuncture as a parameterized input into a layered physiological control system. Arrows indicate forward signal processing; the lower arc indicates feedback-driven adjustment toward homeostasis.

Table 1. Existing meridian theories reinterpreted within the five-layer framework.

Theory or evidence stream	Main contribution	Layer emphasized	Remaining limitation
Fascia/connective tissue	Physical continuity, mechanotransduction and extracellular matrix participation	Hardware	Does not fully explain signal meaning or organ-specific outputs
Electrical impedance studies	Measurable electrical differences at some points and lines	Code	Method heterogeneity and need for stronger blinding/standardization
Somatosensory-autonomic circuits	Causal routes from body-surface stimulation to organ/autonomic effects	Hardware/compiler	Explains selected effects rather than the whole meridian map
Neuroimaging	Central representation and network modulation after acupoint stimulation	Compiler	Mostly correlational and state-sensitive
Molecular/neuroimmune studies	Execution through cytokines, neurotransmitters, endocrine and immune pathways	Execution	Often downstream and not sufficient to identify the input code
Precision acupuncture and AI	Quantified parameters and individualized prediction	State correction	Requires prospective validation and causal inference

Table 1. Existing meridian theories reinterpreted within the five-layer framework.

Theory or evidence stream	Main contribution	Layer emphasized	Remaining limitation
Fascia/connective tissue	Physical continuity, mechanotransduction and extracellular matrix participation	Hardware	Does not fully explain signal meaning or organ-specific outputs
Electrical impedance studies	Measurable electrical differences at some points and lines	Code	Method heterogeneity and need for stronger blinding/standardization
Somatosensory-autonomic circuits	Causal routes from body-surface stimulation to organ/autonomic effects	Hardware/compiler	Explains selected effects rather than the whole meridian map
Neuroimaging	Central representation and network modulation after acupoint stimulation	Compiler	Mostly correlational and state-sensitive
Molecular/neuroimmune studies	Execution through cytokines, neurotransmitters, endocrine and immune pathways	Execution	Often downstream and not sufficient to identify the input code
Precision acupuncture and AI	Quantified parameters and individualized prediction	State correction	Requires prospective validation and causal inference

4. Testable Predictions and Experimental Roadmap

A hypothesis article should be judged by whether it generates falsifiable predictions. The Meridian-Program Hypothesis predicts that acupoints and meridian-related regions should show stimulus- and state-dependent multimodal signatures; that disrupting defined peripheral or central nodes should selectively alter downstream effects; and that different pathological states should require different optimal stimulation parameters.

First, multimodal signal fingerprinting can combine four-electrode impedance spectroscopy, local mechanosensing, microdialysis, biochemical assays, ultraweak photon detection where feasible, fMRI and physiological outputs. If the code layer is meaningful, ST36, PC6 and LR3 should not merely produce generic arousal signatures; they should produce distinguishable feature patterns that predict target physiological outputs better than chance.

Second, causal chain studies should move from association to intervention. Optogenetics, chemogenetics, local pharmacological blockade, sensory-neuron silencing and pathway-specific recording can test whether a peripheral signal is necessary and sufficient for a defined central and organ response. The strongest test would show that blocking a predicted node abolishes a specific acupuncture effect while sparing unrelated responses.

Third, parameter mapping should test the state-correction layer. Factorial or fractional factorial designs can vary frequency, intensity, depth, duration and schedule across models representing peripheral sensing faults, central sensitization and effector-network dysfunction. Machine-learning models should be evaluated prospectively, not only fitted retrospectively.

Table 2. Proposed experimental roadmap.

Experiment	Main question	Core methods	Supportive result	Falsifying or weakening result
Signal fingerprinting	Are acupoint signals measurable, separable and state-dependent?	Impedance spectroscopy, mechanosensing, microdialysis, fMRI, physiological monitoring	Distinct multimodal clusters predict organ-related outputs	Signals are indistinguishable from sham/non-acupoint controls after rigorous blinding
Causal chain testing	How does peripheral input become central and organ output?	Node-specific blockade, optogenetics, fiber photometry, organ physiology	Predicted blockade selectively abolishes the target response	Effects persist after blocking predicted pathways or are nonspecific across pathways
Parameter-response mapping	Do different fault states require different input parameters?	Factorial designs, orthogonal arrays, dose-response modeling, prospective prediction	Optimal frequency/intensity/depth profiles differ by pathology and predict outcomes	One nonspecific parameter profile works equally across unrelated states

Table 2. Proposed experimental roadmap.

Experiment	Main question	Core methods	Supportive result	Falsifying or weakening result
Signal fingerprinting	Are acupoint signals measurable, separable and state-dependent?	Impedance spectroscopy, mechanosensing, microdialysis, fMRI, physiological monitoring	Distinct multimodal clusters predict organ-related outputs	Signals are indistinguishable from sham/non-acupoint controls after rigorous blinding
Causal chain testing	How does peripheral input become central and organ output?	Node-specific blockade, optogenetics, fiber photometry, organ physiology	Predicted blockade selectively abolishes the target response	Effects persist after blocking predicted pathways or are nonspecific across pathways
Parameter-response mapping	Do different fault states require different input parameters?	Factorial designs, orthogonal arrays, dose-response modeling, prospective prediction	Optimal frequency/intensity/depth profiles differ by pathology and predict outcomes	One nonspecific parameter profile works equally across unrelated states

5. Implications

For neuroscience, the hypothesis encourages systematic mapping of body-surface to visceral control routes beyond classical dermatome descriptions. Electroacupuncture studies already show that body region and stimulus intensity can determine autonomic outputs [12,13,14]. A layered model can guide future work toward body-surface, spinal, brainstem and cortical maps with explicit physiological endpoints.

For systems biology, the hypothesis shifts attention from static molecular lists to dynamic signal processing. Molecular changes after acupuncture should be interpreted in relation to input parameters, central state and organ-level output. This may help explain why the same point or prescription sometimes produces different clinical responses.

For precision medicine, the state-correction layer suggests that future acupuncture prescriptions could be expressed as parameter sets rather than only point sets. Quantitative manipulation research, AI-assisted prediction and digital acupuncture platforms may become clinically useful only if they are anchored to falsifiable biological models [23,24,25,26].

6. Limitations

This article proposes a theoretical framework. It should not be read as proof that classical meridians exist as discrete anatomical entities. The available evidence supports components of a distributed control system, but no study has yet demonstrated the complete five-layer process in a single experiment.

Several risks must be acknowledged. First, the program metaphor can be overextended; biological systems are adaptive, noisy and embodied, unlike engineered software. Second, some cited evidence is correlational, and neuroimaging findings require careful control for expectation, sensation and analysis pipelines. Third, acupuncture effects are state-dependent, which makes replication more difficult but also makes mechanistic modeling more important. Fourth, traditional meridian maps contain historical and clinical knowledge that may not map cleanly onto one modern biological layer.

The hypothesis will be useful only if it leads to sharper experiments. It should be revised or rejected where rigorous data contradict its predictions.

7. Conclusion

The Meridian-Program Hypothesis reframes the acupuncture meridian problem from an anatomical search for hidden tubes to a systems question about control, information and feedback. It proposes that acupuncture effects may emerge from interactions among connective-tissue and neural hardware, multimodal biological signals, central interpretation, molecular execution and parameterized feedback intervention. This model is deliberately provisional. Its value lies in making old concepts experimentally addressable: identify the signals, map the routes, test the interpreters, measure the outputs and optimize the parameters. If supported, the framework could provide a common language for traditional meridian theory, neuroscience, systems biology and precision neuromodulation. If falsified, it should still help clarify which parts of the meridian concept belong to measurable physiology and which require different explanations.