Weekly Update — July 21–27, 2026: A Pharmacodiagnostic Compass, Six New Chapters, Companion Papers, and the Dose That Is a Diagnostic
Last week’s update covered genetics, cell-type enrichment, and skeletal asymmetry. It’s hard to communicate the scope of what happened since then, so I’ll structure this by theme rather than timeline. The short version: the paper now contains a comprehensive pharmacodiagnostic framework spanning 102 drugs, a 17-medication hormesis model decomposed into four mechanistically distinct categories, six new chapters (Comparative Nosology, Research Methods, Global Perspectives, Economic Impact), 924 glossary entries with live tooltips on the web version, and the first formal pharmacodiagnostic matrix linking drug response patterns to probable underlying mechanism. The repository saw roughly 213 commits across seven days.
1 The pharmacodiagnostic compass
The largest single thread of work was a pharmacodiagnostic infrastructure that did not exist in any form last Monday.
Chapter 33: Mechanistic Cascade Tracing now contains 102 medication, supplement, and treatment entries — each with a step-by-step cascade trace through the mechanisms a drug engages, the response patterns the drug’s effects and side effects reveal, and what non-response or paradoxical worsening implies about which pathways are not the bottleneck. The chapter includes an 11-mechanism worsening taxonomy for drugs that can provoke PEM or protracted deterioration, a formal Pharmacodiagnostic Matrix that cross-indexes all 102 drugs against 28 hypothesis domains, and dissociation patterns — pairs of drugs that share one target but differ on another, so that differential response localizes the lesion.
A new section on negative-control class drugs defines a formal taxonomy: rituximab as canonical negative control, immunoglobulin adsorption as boundary case, and the “null ladder” vs. “negative-control ladder” distinction that prevents slipping from “null result” into “null claim.” This is both a methodological safeguard — every hypothesis in the paper must specify what result from what class of negative control would falsify it — and a diagnostic instrument: the drug that doesn’t work is information about the mechanism that isn’t present.
Pulsed therapy and time-dependent hormesis are now integrated as a general treatment framework. Some drugs have a therapeutic window that is time-dependent, not just dose-dependent — intermittent dosing preserves selectivity across targets with different off-rates (rapamycin, mTORC1 vs. mTORC2), prevents receptor desensitization that extinguishes hormetic compensation, and exploits the recovery phase during which the adaptive response is higher than baseline. A companion speculation — interoceptive dose-finding — reframes deliberate intermittent dosing as rational N-of-1 experimentation: the patient titrates toward the dose that produces an internal signal of benefit, and the titration pattern itself becomes diagnostic.
The TRPV1-temperature axis is now integrated as an autonomic stress test in the integrative speculations section of Chapter 15. Thermal challenges — cold exposure, hot bath, sauna — differentially stress TRPV1-mediated vasomotor, sudomotor, and baroreflex pathways. Response patterns to these challenges (exaggerated vasoconstriction, absent sweating, delayed heart-rate recovery) localize autonomic failure to specific receptor-channel complexes and predict which pharmacologic probes (capsaicin, menthol, TRPV1 antagonists) are most likely to reveal pathology. A dedicated proposed study now specifies the design.
2 The 17-drug hormesis framework and its four categories
The hormesis framework expanded from 5 to 17 medications, supported by 11 new citations and a 5th mechanistic cluster (D2 partial agonists, separating aripiprazole/LDA from the Nrf2 cluster), now integrated into the Integrative Models chapter. The four categories — and their diagnostic implications — are:
- Category 1: Nrf2-mediated hormesis (LDN, low-dose lithium, melatonin, sulforaphane, NAC, quercetin). The drug triggers a mild stress signal; the cell’s adaptive response via Keap1-Nrf2-ARE is the actual therapeutic mechanism. The therapeutic window reflects Nrf2 transcriptional reserve. Narrow windows predict narrow windows for all cluster drugs.
- Category 2: Partial-agonist inverted-U (LDA/aripiprazole). Not hormesis — a receptor-occupancy property. At low occupancy, net agonism; at higher occupancy, the partial agonist displaces the stronger endogenous ligand and inverts to net antagonism. The inversion point reveals D2 receptor reserve.
- Category 3: Catecholamine inverted-U at PFC (modafinil, duloxetine, beta-blockers, guanfacine). These drugs push along the prefrontal cortex’s native inverted-U dopamine/noradrenaline curve. The dose at which benefit appears, peaks, and inverts reveals baseline catecholamine tone.
- Category 4: Concentration-dependent target selection (rapamycin, corticosteroids, DORAs, H1 antihistamines, allopregnanolone, ketotifen). Each drug’s inverted-U arises from compound-specific biochemistry — differential binding affinity, CNS penetration thresholds, or biphasic metabolism.
The core falsifiable hypothesis: drug inversion points should cluster by category. Nrf2-cluster drugs should have correlated windows. Catecholamine-cluster drugs should have correlated windows. The two clusters should not correlate. This has never been tested, and the paper now specifies a trial design (HIP-B) to test it.
Separately, Chapter 8 now integrates an HSV dormancy-undormancy probe: visible herpesvirus reactivation (cold sores) as a patient-reportable, PEM-linked endogenous reactivation biomarker, exploiting HSV-1/2 latency cycling as a window into the broader herpesvirus reactivation dynamics implicated in ME/CFS symptom flares.
The companion blog series — Why More Isn’t Better, The Inverted-U Is Not One Thing, Your LDN Dose Is a Diagnostic, and The Pharmacopoeia — Every Dose Range for Every Medication and What It Means — is the patient-facing equivalent of the consolidated reference table.
Two final extensions to the pharmacodiagnostic framework arrived late in the week:
Therapeutic-depth classification: all 193 drug entries in sec-12 and all 208 glossary medication entries were classified along a Restorative→Symptomatic axis, systematizing the existing ch32 three-level taxonomy across the full paper. Each drug now carries a depth rating that distinguishes fundamental mechanism repair from symptom palliation — making explicit what the pharmacodiagnostic inference can and cannot conclude from a given response.
Ligand-Receptor Duality: a meta-framework formalizing the observation that most receptor-targeting drugs activate one signaling pathway while suppressing another at the same receptor. Aripiprazole at D2 activates cAMP while suppressing β-arrestin — a biased agonist, not a simple partial agonist. This duality means pharmacodiagnostic inference must decompose “drug X works” into “which arm of the receptor’s signaling portfolio does the patient’s system need?” The framework is now embedded in the sec-12 structure.
3 Transcriptomics: Amatica Health RNA-seq
Preliminary findings from the Amatica Health RNA-seq cohort were integrated into Chapter 6. These are the first ME/CFS transcriptomic data to include patients with confirmed comorbid hypermobility spectrum disorders, creating a natural stratification axis. The findings converge with existing transcriptomic signatures — mitochondrial downregulation, immune dysregulation, and metabolic shift — while surfacing new gene-expression patterns that differ between hypermobile and non-hypermobile strata, suggesting distinct transcriptomic architecture within the ME/CFS umbrella.
4 Muscle microvascular dysfunction and brain clearance
Two structural pathophysiology threads were woven across multiple chapters:
Skeletal muscle capillary microvascular dysfunction — the Slaghekke/Charlton findings on capillary basement membrane thickening and impaired oxygen extraction — was integrated across Chapters 6, 10, 13, 14k, and 17. This is a bridge mechanism: it connects the systemic microvascular pathology to the muscle-specific PEM cascade, providing a concrete structural correlate for the impaired oxygen delivery that makes even minor exertion catastrophic.
Brain clearance architecture — Chapter 13 was populated with a full glymphatic and meningeal lymphatic framework. The model explains why unrefreshing sleep may be mechanistically specific rather than merely symptomatic: impaired glymphatic clearance during sleep means each night fails to clear the metabolic byproducts of the previous day’s neural activity, producing a cumulative neurotoxic load that sleep cannot reset.
5 T-cell mitochondrial exhaustion
The t-cell-mitochondrial-exhaustion topic completed its full integration pipeline — bridging Chapter 6 and Chapter 7, receiving a Phase 7 certainty bump, and adding glossary entries. The framework describes a self-reinforcing loop in which chronic immune activation drives mitochondrial exhaustion in T-cells, exhausted T-cells lose the metabolic capacity to clear persistent antigens, and residual antigen drives further activation. This is not a separate mechanism from the paper’s mitochondrial themes but a cell-type-specific instantiation of the same energy-failure principle — applied to the very cells that should be resolving the immune dysfunction.
6 Companion papers: research ideas and software ideas
Two new standalone companion papers were added:
Research Ideas — 15 formal research proposals spanning diagnostics, treatment trials, biomarker validation, and epidemiological studies, each with specified design, feasibility assessment, and expected evidentiary weight.
Software Ideas — 17 patient-ranked software concepts addressing the practical infrastructure gap: symptom trackers that capture PEM delay, treatment registries with dose-response surfaces, N-of-1 trial platforms, and differential-diagnosis decision support. These are concrete specifications, not feature lists — each includes the data model and inference logic.
7 HSAT2 long abstract
Geneviève’s long abstract for ISLC-PAIS 2026 was finalized and submitted. The abstract presents the HSAT2 case’s microglial sensitization mechanism as a worked example bridging patient-observed pharmacology (LDA dose-response) to formal mechanistic modeling, with an SVG genome-organization figure. It’s the first academic presentation of the paper’s patient-sourced data layer.
8 The Subjective-Measurable Discrepancy Index
Chapter 15 now introduces a formal diagnostic construct in its dedicated section: the gap between what a patient reports they can do and what objective testing measures they can do. In ME/CFS, patients frequently perform normally on single-domain tests — grip strength, VO₂ max, cognitive battery — while being unable to sustain that performance across domains or across days. The Discrepancy Index quantifies this: high discrepancy is diagnostically informative, not evidence of malingering. It reflects the ecological validity gap in standard testing: single-domain, single-session, non-fatigued-state measurement misses the multi-system, effort-contingent, delayed-crash nature of the disease. The framework is now embedded in the integrative-models chapter with a full annotated bibliography and diagnostic threshold specification.
9 Six new chapters
Chapter 16: Comparative Nosology — written from scratch. Systematically compares ME/CFS diagnostic criteria (Fukuda, CCC, ICC, IOM, NICE) across domains, with a load-bearing vs. secondary lock decomposition, epigenetic consolidation model, GPCR autoantibody cascade, and PANS/PANDAS integrated as Architecture B existence proof for neuroimmune encephalitis spectra. The chapter formalizes the distinction between locks that sustain the disease state and locks that amplify it.
Chapter 34: Research Methods — developed from stub. Eight-subtopic methodology analysis covering CPET protocols, biomarker validation frameworks, N-of-1 trial design, Bayesian adaptive designs, negative-control selection criteria, and the pharmacodiagnostic inference logic. The chapter provides the methodological scaffolding for every proposed study in the paper.
Chapter 40: Economic Impact — full cost-of-illness, employment, and burden-to-funding disparity analysis. Documents the systematic mismatch between disease burden (prevalence × disability weight × duration) and research funding, with international comparator data.
Chapter 43: Global Perspectives — 18 countries across 6 dimensions (diagnostic criteria, healthcare access, disability recognition, research infrastructure, patient advocacy, cultural attitudes). The chapter maps the geopolitical landscape of ME/CFS care and research, identifying where treatment protocols diverge and where they converge.
Chapter 39: Healthcare Systems and Policy — 17 treatment environments, 10 papers, and a 4-failure model (diagnostic failure, therapeutic failure, systemic failure, knowledge-production failure) that explains why ME/CFS outcomes are poor across health systems with fundamentally different designs. Structural failure is downstream of category error, not resource allocation.
Chapter 45: Patient-Generated Knowledge and Citizen Science — documents the parallel knowledge-production ecosystem built by patients outside formal institutions: protocol sharing, crowdsourced treatment registries, adverse-event tracking, and the informal diagnostic expertise that experienced patients accumulate.
Additional stubs for 8 further chapters were created to close structural gaps. The entire chapter numbering was reorganized (1–55 sequentially) to accommodate the structural expansion. The existing history of ME/CFS chapter (ch02) was populated with a full chronological narrative from 1934 to 2025, reconciled with overlapping content in ch01. Satellite cell regeneration environments lost in a prior restructuring of Chapter 7 were restored and received adversarial review fixes — a mechanical repair that proved the benefit of the Phase 10b strategic-framing propagation skill. A serotonergic bottleneck integration plan and search log was queued for future integration, alongside formal topic integration priorities and creative integration priorities documents that rank the remaining open topics by estimated impact and difficulty.
10 Web and infrastructure
The glossary system expanded from zero to 924 entries with live tooltips across the web version, plus 172 aliases and a no-tooltip exception mechanism. The system spans medications, supplements, pathogens, anatomy, methods, and domain terminology. Every technical term in the web paper is now a hover target.
The Typst-to-QMD conversion pipeline was refactored into an in-process parallel pipeline with 10x throughput improvement. A new BuildAuditTest runs 10 post-build checks. The deploy timeout was restored to 60 minutes. Duplicate build runs were eliminated. Bib files are now auto-generated via BuildWeb.java rather than tracked in git.
Three LinkedIn posts were naturalized in prose style. The KCE Belgian ME/CFS needs assessment and a 4-part DecodeME GWAS series were published.
11 Patient-sourced data
Two clinical observations were logged: a Libramont PEM episode triggered by LDA dose reduction from 1 mg to 0.5 mg (suggesting 0.5 mg is below the microglial D2 engagement threshold for this patient), and LDA fault tolerance — the observation that sleep deprivation on LDA no longer triggers migraine or brain fog, with a manic-switch differential that contrasts with prior stimulant excitation patterns. The HSAT2 case’s microglial sensitization mechanism was updated for the ISLC-PAIS 2026 long abstract submission.
12 What is now falsifiable that was not last week
The paper’s falsifiability surface expanded substantially. The Pharmacodiagnostic Matrix specifies exact cross-drug inference patterns that would be contradicted by specific response-nonresponse pairings. The four-category hormesis decomposition predicts within-category correlation of drug windows and between-category independence — a claim testable in any cohort with multi-drug dose-response data. The Subjective-Measurable Discrepancy Index specifies a threshold above which discrepancy counts as diagnostic. The negative-control ladder defines which classes of null result falsify which hypotheses. Every new chapter — global perspectives, economic impact, healthcare systems — carries its own falsifiable predictions in the Phase 5a review records.
The hypothesis registry was reorganized by topic, and all entries now carry falsifiability conditions, not just existence claims. The T-cell mitochondrial exhaustion model is falsifiable via single-cell metabolomics: exhausted ME/CFS T-cells should show lower spare respiratory capacity than healthy controls, and the deficit should correlate with antigen-specific tetramer staining — if not, the energy-failure explanation for persistent viral/autoimmune targets is wrong. The glymphatic framework is falsifiable via contrast-enhanced MRI: if ME/CFS patients show normal glymphatic clearance rates, unrefreshing sleep must be driven by a different mechanism.
This update covers work committed between July 21 and July 27, 2026. The full changelog and paper are at loth.cc. Previous weekly updates: Load-Bearing Walls, Root Cause ≠ Treatment Priority, The Genetics That Ends the Debate.