Skip to content

Research hypothesis

The Atlas Hypothesis

ME/CFS as a phenotype-informed failure of biological stress resolution

Research hypothesisNot clinical guidanceUnvalidated modelRequires prospective testing
Status
Research hypothesis
Evidence status
Hypothesis-generating synthesis
Clinical status
Not diagnostic guidance. Not treatment guidance.
Main limitation
The temporal sequence from exertion to PEM is not yet established.

Read this first

This page is highly speculative. It is published as brainstorming material and as a starting point for dialog with researchers, clinicians and patient advocates — not as a finding, not as a model the Atlas endorses, and not as guidance for any decision a reader might face. The cautious phrasing throughout (“may”, “could”, “candidate”, “proposed”) is the appropriate epistemic register. Nothing below should be read as established.

Abstract

Post-exertional malaise (PEM) is the defining clinical feature of ME/CFS, but the biological sequence that produces delayed relapse remains unresolved. Current research describes abnormalities across immune function, cellular energy metabolism, autonomic regulation, perfusion physiology and CNS function. These abnormalities are repeatedly discussed across the ME/CFS literature, but most studies measure one system at a time, often at rest. The field lacks a prospective 0–72 hour post-exertional study that measures metabolic, immune, autonomic, perfusion and CNS changes in the same individuals while tracking PEM onset and recovery.

The Atlas Hypothesis proposes a research framework: ME/CFS may involve a persistent baseline vulnerability across immune, metabolic, autonomic, perfusion and CNS systems. In some patients, this vulnerability may follow infection-triggered immune reprogramming and possible regulatory or epigenetic locking. Exertion, cognitive demand, orthostatic stress or sensory stress may then exceed the system’s reduced recovery capacity. PEM emerges when biological stress resolution fails.

This model does not propose one universal linear mechanism. It proposes phenotype-informed convergence: different patients may reach the same clinical endpoint, delayed PEM, through different dominant biological pathways [murovska-2026-myalgic-encephalomyelitis].

1. Why this hypothesis exists

ME/CFS research often asks:

What is abnormal at baseline?

That question is necessary, but insufficient. PEM is a dynamic phenomenon. A patient may appear similar to baseline immediately after activity, then deteriorate hours or days later.

The Atlas Hypothesis reframes the central question:

What fails to normalize after biological stress?

The hypothesis therefore focuses on recovery, resolution and reset rather than only static abnormality.

2. Core claim

ME/CFS may be best studied as a phenotype-informed failure of biological stress resolution in a vulnerable immune-metabolic-autonomic-CNS network.

This is not a proven mechanism. It is a testable research framework.

3. Model overview

The proposed sequence from biological trigger to delayed PEM, with multiple candidate phenotype pathways converging on the same clinical endpoint.

Biological trigger (infection or other stressor)
Possible immune or regulatory reprogramming
Baseline vulnerability platform
Physical, cognitive, orthostatic, sensory or infectious stress
Reduced recovery capacity exceeded
Failed biological stress resolution
Candidate phenotype pathways (see §6)
Delayed PEM

Candidate phenotype pathways: autonomic / preload / perfusion; immune-reset; metabolic-recovery; CNS / neuroinflammatory; and gut-vagal, respiratory or vestibular modifiers.

4. Baseline vulnerability versus exertion-triggered cascade

LayerWhat it meansEvidence status
Baseline vulnerabilityPersistent abnormalities or lowered reserve before exertionSupported across several domains, but mechanisms vary
Stress exposurePhysical, cognitive, orthostatic, sensory or infectious loadClinically central to PEM
Failed resolutionIncomplete return to baseline after stressStrongly plausible, but not directly mapped across systems
PEM endpointDelayed relapse in function and symptomsClinically established, but biological timing remains unresolved

The model separates baseline disease state from the PEM cascade. A patient can have baseline abnormalities without those abnormalities alone explaining the delayed crash.

5. Why PEM may be delayed

A purely immediate energy-depletion model is incomplete, because PEM often appears after delay. A purely delayed immune model is also incomplete, because early autonomic and metabolic responses may occur before symptoms peak.

The Atlas Hypothesis proposes a staged but not yet proven sequence:

0–4 hours

Acute compensation, sympathetic activation, early metabolic strain

Inferred

4–12 hours

Possible purinergic / DAMP signaling, immune signaling, failed early reset

Inferred

12–72 hours

Clinical PEM emerges as recovery failure becomes systemic

Observed clinically; mechanism unresolved

No current evidence base establishes this full sequence as fact.

6. Candidate phenotype pathways

The model does not assume that all patients crash through the same pathway.

Candidate pathwayPhenotype patternEvidenceMain uncertainty
Autonomic / preload / perfusion dominantOrthostatic intolerance, exaggerated heart-rate rise, low perfusion reserve, cognitive worsening upright [schiweck-2026-systematic-literature]ModeratePrimary driver versus compensatory failure unclear
Immune-reset candidatePost-infectious onset, immune activation, possible viral or autoimmune overlapModerateSerial post-exertional immune recovery data missing
Metabolic-recovery candidateSevere exertional intolerance, prolonged recovery, metabolic abnormalities [naviaux-2016-metabolomics]ModerateEarly ATP / lactate kinetics not established
CNS / neuroinflammatory candidateBrain fog, sensory overload, cognitive deterioration [penson-2026-evaluating-working]Low to moderateTiming and causality remain unclear
Gut-vagal modifierGI symptoms, dysbiosis, vagal / autonomic instability [kim-2026-potential-application]LowCausality and subgroup boundaries unclear
Respiratory / ventilation modifierDyspnea, breathing-pattern disorder, ventilation-perfusion mismatch [wells-2026-breath-mind]Low to moderateRelationship to PEM unclear
Vestibular / sensory modifierDizziness, balance problems, sensory overload, proprioceptive instability [sirotiak-2026-understanding-concussion]Low to moderateMay be amplifier rather than core PEM pathway

These are candidate phenotype pathways, not validated biological subtypes.

Phenotype convergence

Autonomic / preload / perfusion
Immune-reset
Metabolic-recovery
CNS / neuroinflammatory
Gut-vagal modifier
Respiratory / vestibular modifier
Same clinical endpoint: delayed PEM

Different patients may converge on delayed PEM through different dominant pathways. The pathway labels are hypothesis-generating, not validated subtypes.

7. Individual phenotype example

A phenotype with post-infectious onset, low resting heart rate, exaggerated heart-rate rise with light activity, orthostatic or volume intolerance, delayed PEM, cognitive worsening and sensory sensitivity may map most strongly to an autonomic / preload / perfusion-dominant PEM pattern.

This is a phenotype match, not a diagnosis, not a treatment recommendation, and not proof of mechanism.

In this pattern, autonomic / preload / perfusion failure may be the most visible clinical bottleneck, while immune-metabolic dysregulation may be a deeper vulnerability.

Post-infectious vulnerability
Immune-metabolic instability
CNS / autonomic network vulnerability
Low preload or perfusion reserve
Stress exposure
Compensation fails
Delayed PEM

8. Evidence map

8.1 Post-exertional functional decline

Two-day CPET findings are among the strongest objective anchors for PEM because they show reduced function after exertional challenge rather than only baseline impairment. This supports the idea of failed recovery, but does not identify the upstream mechanism.

8.2 Metabolic and energy-related findings

Recent multi-system studies report abnormalities in energy metabolism, including altered AMP, ADP, ATP/ADP ratio, NAD-related metabolism and kynurenine pathway markers [naviaux-2016-metabolomics]. These findings support metabolic involvement, but do not prove that ATP failure is the first event after exertion.

8.3 Immune findings

ME/CFS research repeatedly reports immune abnormalities, including NK-cell and T-cell related findings in several cohorts. These findings support immune vulnerability, but serial post-exertional immune-reset data are still missing.

8.4 Autonomic and perfusion findings

Orthostatic intolerance, POTS-like physiology, preload limitations and cerebral blood-flow abnormalities are relevant to PEM, particularly in autonomic / perfusion-dominant phenotypes [schiweck-2026-systematic-literature]. These mechanisms may be primary in some patients and compensatory or downstream in others.

8.5 CNS findings

Cognitive dysfunction, sensory sensitivity and limited neuroinflammatory findings support CNS involvement [penson-2026-evaluating-working]. The model therefore treats CNS / autonomic regulation as a central node, but not as proven primary cause.

8.6 Possible regulatory or epigenetic locking

Epigenetic findings, including OPRM1-related observations in the Atlas corpus [wyns-2026-hypermethylation-oprm1], may suggest stable regulatory changes. These findings should be treated as candidate signals, not proof of causality.

9. What the model explains

9.1 Normal routine resting labs

The model explains why routine resting tests can be normal: the key abnormality may lie in dynamic recovery after stress, not in static resting measurements.

9.2 Delayed PEM

The model explains delayed PEM as failed stress resolution after initial compensation.

9.3 Heterogeneity

The model explains heterogeneity by allowing different dominant pathways to converge on the same clinical endpoint.

9.4 Chronicity

The model allows for persistent disease through immune, metabolic, autonomic or regulatory set-point changes, without assuming one universal cause.

10. What the model does not prove

This hypothesis does not prove that:

  • ATP failure is the first event after exertion
  • OPRM1 or any single epigenetic marker causes ME/CFS
  • immune reset failure is present in every patient
  • autonomic dysfunction is always primary
  • neuroinflammation is the main driver in all patients
  • all ME/CFS patients share one pathway
  • any treatment follows from this framework

This is a research model, not clinical guidance.

11. Main weaknesses

WeaknessWhy it matters
No prospective 0–72 hour serial PEM studyThe temporal sequence is inferred
No direct early ATP / lactate kineticsThe first biological domino remains unknown
No serial immune-reset dataFailed immune resolution remains plausible, not established
Autonomic findings may be compensatoryAutonomic dysfunction may be bridge or consequence
CNS findings may be focalDiffuse hypoperfusion cannot explain all cognitive patterns
Candidate pathways are not validated subtypesStratification remains hypothesis-generating
Long COVID and ME/CFS overlap but are not identicalPost-COVID findings cannot be automatically generalized [soares-2026-recommended-long]

12. What would falsify or weaken the hypothesis

FindingImpact
PEM occurs without measurable recovery failure in any tested biological systemWeakens failed-resolution framework
Baseline phenotype does not predict post-exertional pathway or PEM patternWeakens phenotype-informed convergence
Serial metabolic, immune, autonomic and CNS markers normalize before PEM beginsWeakens delayed biological cascade
A single universal mechanism explains most patients across phenotypesWeakens subtype / convergence model
Candidate epigenetic or regulatory markers are inconsistent and unrelated to courseWeakens locking component

13. Decisive study needed

The decisive next study is not another isolated baseline comparison. It is a prospective, phenotype-stratified, 0–72 hour PEM study.

ComponentRequirement
CohortStrictly defined ME/CFS with documented PEM, plus matched sedentary controls
Baseline stratificationImmune, metabolic, autonomic, CNS / perfusion and symptom phenotype
Stress triggerStandardized exertional, orthostatic or cognitive challenge
TimepointsBaseline, immediate, 4h, 12h, 24h, 48h, 72h
Metabolic markersLactate, pyruvate, acylcarnitines, ATP / ADP-related markers, NAD-related markers
Immune markersNK function, cytokines, chemokines, immune-cell phenotype
Autonomic markersHRV, beat-to-beat blood pressure, orthostatic response, preload indicators
Perfusion / CNSCerebral blood flow by TCD, NIRS or imaging
OutcomePEM onset, severity, duration and recovery
Primary testWhether baseline phenotype predicts post-exertional collapse pathway

14. Final formulation

The Atlas Hypothesis proposes that ME/CFS may be best studied as a phenotype-informed failure of biological stress resolution. A persistent baseline vulnerability across immune, metabolic, autonomic, perfusion and CNS systems may lower the capacity to recover after stress. PEM may emerge when exertion exceeds this reduced reset capacity and the patient collapses through a dominant biological pathway.

The Atlas Hypothesis should be tested, not assumed.

References cited on this page

Each link opens the Atlas evidence page for the underlying study, where the primary paper, classification, and machine-summary disclosure are visible. Citations are illustrative for the framework above; this page does not re-summarise the primary findings of these studies.

Related Atlas pages

  • Methodology — how the Atlas is built, version stamps, limitations.
  • Evidence Atlas — searchable index of the underlying studies.
  • Glossary — definitions for PEM, evidence levels, and other terms used above.
  • Understanding PEM — Atlas guide to post-exertional malaise.
  • FAQ

This page is highly speculative. It is published as brainstorming material and as a starting point for dialog — not as a finding, not as a model the Atlas endorses, not as a consensus statement, and not as diagnostic or treatment guidance. Cautious language is intentional: the model should be tested, not assumed.