{
  "slug": "gut-brain-psychiatric",
  "round": 2,
  "date": "2026-10-14",
  "authoring_model": "Sol 6.1",
  "reviewer": "Claude Opus 5.5 (Anthropic)",
  "level_before": "Γ",
  "level_after": "Γ",
  "confidence_before": "Moderate",
  "confidence_after": "Moderate",
  "role_before": "reinforcing / stabilizing factor",
  "role_after": "reinforcing / stabilizing factor",
  "pmids_checked": 23,
  "pmids_failed": 0,
  "claims_spot_checked": 16,
  "validator": "PASS (34 checks, 0 FAIL, 0 WARN)",
  "findings": [
    {
      "id": "G-1",
      "severity": "Major",
      "location": "Consensus provenance reconciliation - 9 October 2026",
      "original": "The earlier adversarial approval refers to its original edition",
      "corrected": "No earlier adversarial approval exists for this case",
      "evidence": "no approval section/banner/line for this slug anywhere in deliverables; regression introduced by the Oct 9 revision",
      "type": "false_approval_reference"
    },
    {
      "id": "G-2",
      "severity": "Editorial",
      "location": "Human controlled counterevidence and the pathway gap",
      "original": "did not outperform autologous FMT on MADRS (16.74 versus 15.4 points, P=.64)",
      "corrected": "stated explicitly as baseline-to-week-24 improvements, with 95% CI -7.3 to 4.6 added",
      "evidence": "PMID 41785480: 16.74-point (allogenic) versus 15.4-point (autologous) improvement, t=-0.47, P=0.64. As printed the numbers could be read as endpoint MADRS scores; this is the ledger's most load-bearing negative result",
      "type": "numeric_readability"
    },
    {
      "id": "G-3",
      "severity": "Minor",
      "location": "Human controlled counterevidence and the pathway gap",
      "original": "several morning heart-rate-variability measures were higher in the probiotic arm after three months",
      "corrected": "abstract comparison named as probiotic-treated patients versus healthy controls; randomised within-patient contrast cited from the body (morning RMSSD U=241.00, Z=3.03, P=.002; morning lnHF t(33)=-3.48, P=.001, d=-1.19) plus the prespecified primary endpoint",
      "evidence": "PMID 40298641 abstract comparator is healthy controls, which confounds intervention with diagnosis and cannot establish a treatment effect; PMC12045568 body reports the randomised within-MD contrasts, so the review's dissociation inference survives on better support. Not load-bearing for a subclaim level: SC4's level rests on the absent human mediator experiment and the allogenic-versus-autologous contrast",
      "type": "comparator_misstatement"
    }
  ],
  "rejected_findings": [
    {
      "item": "probiotic and FMT efficacy statements",
      "reason": "all match sources including primary-versus-secondary and ITT-versus-completer distinctions (PRO-DEMET, Borkent, Romijn, Kazemi, Elahinejad, Schaub, Pinto-Sanchez, Nikolova, Green, Moerkl)"
    },
    {
      "item": "Li FMT trial reading",
      "reason": "abstract confirms null week-8 remission and greater HAMD-17 reduction; body-level figures paywalled and not independently re-verifiable, recorded as residual limitation; the review's disagreement with the abstract's 'durable engraftment' is raised openly"
    },
    {
      "item": "confidence band",
      "reason": "rationale already rests on evidence density and reproducibility with no appeal to access depth; Moderate stands under the 2026-10-14 lead policy"
    },
    {
      "item": "scope-matched aggregation and departure from the index weakest-link rule",
      "reason": "SC4 is the scope-matched subclaim; narrower subclaims reported separately; all three rules reported and the index answer carried as explicit dissent"
    },
    {
      "item": "Best-supported rival renderer artifact",
      "reason": "rank 1 is a non-nominal model; the nominal M3 ranks 3 of 5"
    },
    {
      "item": "momentum / stability values",
      "reason": "both carry the required artifact; no 'Decisive work underway' claim"
    },
    {
      "item": "collective-author defects (both variants)",
      "reason": "no dropped author with stray comma and no consortium name truncated to a surname fragment; no record in this ledger has a collective first author"
    },
    {
      "item": "rendering defects",
      "reason": "none present"
    },
    {
      "item": "lossy PubMed abstract fields (PMID 40298641)",
      "reason": "abstract garbled mid-sentence with the taxon name dropped ('a significant increase in , particularly along with improved sleep'; full text gives Christensenellales, particularly Akkermansia muciniphila) - same lossy-field pattern as the dropped-collective-author defect; no text in this review depends on it"
    }
  ],
  "propagate_to": [],
  "divergent_copies": [],
  "approval": "Approved as revised"
}