Formal Model Proposals for Causal Hierarchy Validation
The following proposals formalize specific mechanisms from the causal hierarchy as ordinary differential equation (ODE) systems. These models are designed to be instantiated in Chapter Formal Causal Hierarchy Analysis and tested against empirical data.
1 SPM Resolution ODE Model
Proposal: Introduce a new dynamic variable SPM(t) representing the systemic concentration of specialized pro-resolving mediators (resolvins, protectins, maresins, lipoxins). The rate equation is:
\[(\text{dSPM})/(\text{d}t) = k_\text{sub} S_\text{substrate} - d \text{SPM} - k_\text{pem} \text{PEM}(t)\]
where \(S_\text{substrate}\) is omega-3 precursor availability (EPA/DHA), \(d\) is the degradation rate constant, and \(k_\text{pem}\) represents exertion-dependent SPM consumption or inactivation during PEM episodes. The PEM(t) term couples SPM dynamics to the exertion cascade, creating a scenario where repeated exertion before SPM recovery produces cumulative resolution deficit.
Key predictions:
- SPM depletion below a critical threshold \(\theta_\text{SPM}\) prevents normal resolution of post-exertional inflammation, prolonging the recovery window
- tVNS enhances \(k_\text{sub}\) (vagal-SPM coupling via \(\alpha\) 7-nAChR), shifting the steady state upward
- Aspirin-acetylated COX-2 generates aspirin-triggered resolvins (AT-RvD1-4, AT-RvE1-3), providing a bypass mechanism when native SPM synthesis is impaired
- The model predicts a bifurcation: below a critical \(k_\text{sub}\)/\(d\) ratio, the SPM system cannot maintain a healthy steady state, and resolution failure becomes self-sustaining
Testable: SPM lipidomics at multiple time points post-CPET in ME/CFS vs controls would parameterize \(k_\text{sub}\), \(d\), and \(k_\text{pem}\) for each patient. (Proposal only; no ME/CFS SPM data exist.)
Cross-reference: SPM deficiency hypothesis at Family 20: Inflammation Resolution and Lipid Mediators. Cholinergic-SPM coupling at Chapter Neurological and Neurocognitive Dysfunction.
2 HMGB1-Ferroptosis ODE Coupling
Proposal: Model the bidirectional amplification between HMGB1 release and ferroptosis susceptibility as a coupled ODE system. Let \(H(t)\) = extracellular HMGB1 concentration (disulfide + reduced isoforms) and \(F(t)\) = ferroptosis potential (proxied by lipid peroxidation burden).
\[\text{dH}\text{/dt} = \alpha_H dot F - \beta_H dot H\] \[\text{dF}\text{/dt} = \alpha_F dot H - \beta_F dot F + \gamma(t)\]
where \(\alpha_H\) is the rate of HMGB1 release from ferroptotic cells (DAMP release), \(\beta_H\) is HMGB1 clearance, \(\alpha_F\) is HMGB1-driven ferroptosis promotion (via TLR4/RAGE → ROS → lipid peroxidation), \(\beta_F\) is endogenous antioxidant/GPX4-dependent repair, and \(\gamma(t)\) captures exertion-driven oxidative stress.
Key predictions:
- The system admits a stable low (healthy) and a stable high (disease) steady state when \(\alpha_H \alpha_F / (\beta_H \beta_F) > 1\)
- Exercise shifts \(\gamma(t)\), providing a transient perturbation that can trigger state transition in vulnerable individuals
- HMGB1 Box A antagonists reduce \(\alpha_H \alpha_F\), shrinking the disease basin of attraction
- Iron chelation (desferrioxamine) reduces \(F\) by limiting Fenton chemistry, lowering \(\alpha_F\)
- Antioxidants (NAC, vitamin E) increase \(\beta_F\), facilitating return to healthy steady state
Testable: Simultaneous HMGB1 isoform measurement and lipid peroxidation markers (4-HNE, MDA) at rest and post-exertion would provide initial parameter estimates. (Proposal only; no coupled HMGB1-ferroptosis data exist in ME/CFS.)
Cross-reference: HMGB1 DAMP hypothesis at Family 20: Inflammation Resolution and Lipid Mediators. Ferroptosis susceptibility at Ferroptosis Susceptibility.
3 IRE1alpha Bistable Switch Bifurcation Model
Proposal: Model the unfolded protein response (UPR) sensor IRE1alpha as a bistable switch governing the transition from adaptive UPR to chronic inflammatory signaling. Let \(U(t)\) = unfolded protein load in the ER and \(I(t)\) = IRE1alpha activity (spliced XBP1). The core dynamics follow:
\[\text{dI}\text{/dt} = k_I dot U / (K_m + U) - d_I dot I + \sigma(I)\] \[\text{dU}\text{/dt} = p_U - d_U dot U - k_I dot U / (K_m + U)\]
where \(\sigma(I)\) represents autocatalytic IRE1alpha activation (clustering-induced trans-autophosphorylation at high unfolded protein load), \(p_U\) is the baseline protein misfolding rate, and \(d_U\) is ERAD-dependent unfolded protein clearance.
Key predictions:
- At low \(p_U\) (normal physiology), the system has a single stable fixed point (low I, low U — adaptive UPR)
- At high \(p_U\) (oxidative stress, viral infection, mitochondrial dysfunction), the system undergoes a saddle-node bifurcation, gaining a second stable fixed point (high I, high U — chronic UPR with IL-6/TNF-\(\alpha\) production via IRE1alpha-XBP1)
- The bistable switch creates hysteresis: once activated, chronic UPR persists even after \(p_U\) returns to normal
- TUDCA (chemical chaperone) reduces \(U\) directly, potentially collapsing the chronic high-I state
- PERK-eIF2alpha and ATF6 arms provide parallel but slower response channels, setting the timescale for resolution vs. commitment to apoptosis
Testable: Measurement of spliced XBP1, phospho-IRE1alpha, GRP78, and CHOP in ME/CFS PBMCs at baseline and post-exertion would test whether UPR activation follows bistable kinetics. (Proposal only; no IRE1alpha/UPR data exist in ME/CFS.)
Cross-reference: UPR/ER stress section at Family 13: Protein Homeostasis and Degradation. PERK-ATF4-CHOP apoptosis at Chapter Energy Metabolism and Mitochondrial Function.
4 Glutamate-GABA E/I Balance ODE Module
Proposal: Extend the existing ODE family (Watton 2026, Fourel) with a glutamate-GABA E/I balance module. Let \(G(t)\) = extracellular glutamate concentration, \(B(t)\) = extracellular GABA, and \(R_\text{EI}(t)\) = E/I balance ratio. The minimal system is:
\[(\text{d G})/(\text{d}t) = k_\text{rel} - k_\text{uptake} G - k_\text{GAD} G dot P\] \[(\text{d B})/(\text{d}t) = k_\text{GAD} G dot P - k_\text{clear} B\] \[R_\text{EI} = G / (k_\text{half} + B)\]
where \(k_\text{rel}\) = vesicular glutamate release (ATP-dependent), \(k_\text{uptake}\) = astrocytic EAAT2 clearance (ATP-dependent), \(k_\text{GAD}\) = GAD activity converting Glu to GABA (P5P-dependent), \(P\) = available pyridoxal-5’-phosphate, and \(k_\text{clear}\) = GABA clearance. The ATP dependence of both \(k_\text{rel}\) and \(k_\text{uptake}\) creates an asymmetry: EAAT2 requires 3 Na+ per glutamate transported → Na+/K+-ATPase → ~1 ATP per glutamate, while vesicular release requires ~0.5 ATP per glutamate via VGLUT. When ATP drops below a threshold, \(k_\text{uptake}\) decreases faster than \(k_\text{rel}\), creating a “glutamate trap”: \(G\) rises unchecked, shifting \(R_\text{EI}\) toward excitotoxicity.
Key predictions:
- The system shows bistability when genetically reduced clearance (risk variants in astrocytic EAAT2, reduced GAD activity from cofactor depletion) is combined with risk-variant-increased release: two stable steady states (low-Glu healthy, high-Glu excitotoxic) separated by an unstable threshold
- PEM duration in ME/CFS vs controls is predicted by the time required for ATP-dependent clearance to restore \(G\) below the excitotoxicity threshold after exertion-driven glutamate release
- P5P supplementation shifts the bifurcation point by increasing \(k_\text{GAD}\), raising the ATP threshold at which the trap activates
- The genetic PRS parameter shifts the glutamate release setpoint (higher \(k_\text{rel}\) in high-PRS), lowering the ATP threshold at which the trap activates
Cross-reference: Glutamatergic hypothesis Glutamatergic Synaptic Dysfunction as Genetically-Driven Core Mechanism (Chapter Genetic and Epigenetic Factors); P5P-gutamate connection at Section DecodeME Brain Tissue Enrichment.
5 Brain-to-Periphery Efferent Coupling DAG Model
Proposal: Extend the existing causal DAG (Section Pregnenolone Sulfate as TRPM3 Sensitizer) with a formal efferent coupling pathway from CNS genetic liability to peripheral pathology. The canonical path is:
\[\text{CNS Genetic Liability} -> \text{Efferent Coupling} -> \text{Peripheral Pathology}\]
where “Efferent Coupling” is a composite node representing autonomic (sympathetic/parasympathetic), neuroendocrine (HPA axis), and descending pain modulation pathways. This formalizes the brain-first model (Hypothesis Brain-First Genetic Model, Chapter Genetic and Epigenetic Factors) in the graphical causal framework: primary genetic liability operates through neuronal circuits, and peripheral manifestations arise as downstream consequences of efferent signaling dysfunction.
Key predictions:
- The efferent coupling parameter \(\theta_\text{EC} \in [0,1]\) determines the gain from CNS liability to peripheral pathology. At high \(\theta_\text{EC}\) (greater than 0.7), identical genetic liability produces severe peripheral manifestations (orthostatic intolerance, GI dysmotility, immune dysregulation). At low \(\theta_\text{EC}\) (under 0.3), peripheral pathology is minimal despite equivalent CNS liability.
- Causal mediation analysis will show that central glutamate measures (MRS, fMRI) mediate the relationship between glutamatergic PRS and peripheral symptoms (orthostatic intolerance, GI symptoms). The direct path from PRS to peripheral symptoms should be non-significant when central glutamate is included as mediator.
- Patients with high peripheral symptom burden but moderate genetic liability should have high \(\theta_\text{EC}\) — i.e., their efferent coupling amplifies a modest CNS signal into substantial peripheral pathology, explaining the brain-periphery discordance observed in some ME/CFS patients.
Cross-reference: Brain-first genetic model Brain-First Genetic Model; entry points discussion Entry Points: Multiple Doors, One Final Common Pathway; Maccallini 2026 tissue enrichment Maccallini Meta-GWAS: Brain-First Genetic Architecture.
Not all mechanisms are equal. The pathophysiology of ME/CFS, sprawling as it is across immune, neurological, metabolic, cardiovascular, endocrine, and gastrointestinal systems, admits a hierarchy when examined through the lens of causal sufficiency. Four mechanisms—CNS energy crisis, metabolic safe mode lock, GPCR autoantibody cascade, and TRPM3 channelopathy—meet the stringent criteria for trigger-capable root causes: each can account for the full syndrome’s initiation from a healthy baseline, each explains post-exertional malaise, each can sustain itself chronically, and each has a plausible path to multi-system involvement. Six amplifier mechanisms worsen and perpetuate the disease, with epigenetic consolidation and autoimmune persistence identified as load-bearing locks whose removal is necessary for recovery. And a set of downstream consequences, while generating much of the daily symptom burden that defines patients’ lived experience, cannot be addressed in isolation with any expectation of cure.
The hierarchy resolves several longstanding puzzles. Why do different triggering events—viral infection, stress, surgery, vaccination—produce the same disease? Because multiple entry doors converge on a single final common pathway of CNS energy failure and broken restoration machinery. Why do single-target treatments consistently disappoint, even when they target well-validated mechanisms? Because the multi-lock architecture requires simultaneous intervention across mechanisms, and addressing one lock allows the others to compensate. Why does disease duration predict prognosis so strongly, more strongly than initial severity? Because epigenetic consolidation progressively deepens the disease attractor over time, converting a reversible functional state into an entrenched structural one.
The practical implication is that treatment priority should be guided by the product of causal importance and therapeutic tractability—not by causal importance alone. The most upstream mechanism is not necessarily the best therapeutic target if it cannot be reached with current tools. The most tractable intervention is not necessarily the most causally important if it addresses only a secondary lock. Rational treatment strategy navigates between these two axes, combining tractable amplifier-targeting interventions that provide immediate relief with root-cause-targeting interventions that address the disease architecture.
Beyond the core hierarchy, several extensions developed in this chapter and its formal companion (Chapter Formal Causal Hierarchy Analysis) open new directions: the subthreshold reservoir population (Section Multiple Entry Points, Single Final Common Pathway) suggests a preventive medicine approach to ME/CFS, applicable to the large population sitting near the disease separatrix; the threat signal miscalibration concept (Section Metabolic Safe Mode as Trigger-Capable Root Cause) identifies genetic predisposition factors that lower the safe mode activation threshold; and the gut dysbiosis reanalysis (Section Amplifier Mechanisms) challenges the amplifier classification for a specific patient subgroup. The timed passive epigenetic reversal strategy (Section Treatment Priority \(\neq\) Causal Priority) offers a patient-accessible approach to the most intractable lock in the disease, while the TRPM3 sensitization concept explores pharmacological modulation of a recently validated root cause mechanism.
This chapter’s claims are themselves hypotheses, informed by the evidence accumulated in Chapters Energy Metabolism and Mitochondrial Function–Speculative Mechanistic Hypotheses but not proven by it. The sister chapter in Part V (Chapter Formal Causal Hierarchy Analysis) subjects them to formal mathematical testing, encoding the biological reasoning developed here into dynamical systems models that can be interrogated with precision. Where the two approaches agree, confidence increases; where they disagree, the disagreement identifies specific gaps in understanding that future research must address.
The hierarchy is a framework for thinking, not a dogma for believing. It will be revised as evidence accumulates, as new mechanisms are discovered, and as therapeutic experiments test its predictions. The honest acknowledgment of this uncertainty is not a weakness of the framework—it is its defining feature. In a field long plagued by premature certainty, from those who dismissed the illness as psychological to those who promoted specific biological theories without adequate evidence, the most useful contribution may be not the answers, but a disciplined way of asking the questions.