m625 is not a clinical decision tool · not a publisher · not an expert authority · not a simulator. It is a cross-domain Roll-formalized evidence atlas with public audit + attack surface.
Why a patient advocate would come to evidence.x1000.ai
A patient advocate working in chronic-pain support (post-injury, fibromyalgia, post-surgical, or mixed mind-body) reads about brain-network attractor dynamics in lay press and on forums. Pain points (see 05_docs/M625_PERSONAS.md):
it is hard to tell apart cu-prefixed marketing from grounded evidence;
trial information is scattered and the trust signal is unclear.
m625 does not advise this advocate or this advocate's community. **m625 shows what a Roll-formalized evidence pattern looks like, and how to tell it apart from marketing**. The advocate is invited to take the formalization back to the community, audit it independently, and tell m625 where the reading layer is misleading.
This is not patient-facing medical advice. This is not a community-treatment plan. (See "What this demo will not say" below.)
The Roll-formalized object (anchored on m517 CARD01)
m625 does not modify m517's card file. The reference path is /data/projects/m517_NeuroDynamics_Circle/workspace/02_cards_cooked/CARD01_DEEP_v0.2.md (m517 sovereign). What m625 adds here is a patient-advocate-facing reading layer.
How to read this evidence (advocate orientation, ≤ 5 minutes)
Brain-network attractor dynamics is the population-level claim that the brain's resting-state network in chronic-pain populations occupies a different basin of attraction than in pain-free controls. It is not a per-person diagnosis or per-person trajectory.
The viable set V (Kernel §3.2) is the set of network-state trajectories that the cooked evidence supports as accessible across the chronic-pain population.
K4.1 (1, 1, 0) declares: the X-structure (brain network) is visibly altered (b_X = 1); the Φ dynamics is visibly altered (b_Φ = 1); the Δ-B intervention boundary is not altered in the cooked Tu 2020 frame (b_Δ_B = 0). This means: this CARD describes the pattern, not the intervention that flips it.
m517's W7-N hard walls (7 detectors) and NCNU-N 5禁令 ensure no card crosses into individual-patient prediction or care-decision territory. m625 inherits these walls.
What this demo will not say (anti-drift · per 01_materials_cache/C2_m625_anti_drift_manifesto.md)
❌ It does not promise restoration, eradication, or rescue (red-team smuggle pattern AS-001 — cure / reverse lexicon, audit gate G2.2).
❌ It does not assign a per-person Roll instance-mirror object (red-team smuggle pattern AS-006 — digital twin lexicon, audit gate G2.2).
❌ It does not run marketing copy (no limited time, no act now, no 100% effective, no guaranteed — red-team smuggle pattern AS-010, audit gate G2.2 + W6.2 manipulation).
❌ It does not assert collaboration with any named neuroscientist or institution (m517 W7-N.7 — m625 inherits the m517 expert list, audit gate G2.3).
If an advocate reads any of the above into this demo, that is a failure of this demo and should be filed under `_ops/BLOCKED_LOG.md` for next-round revision.
The TMU-A pattern with K4.1 = (1, 1, 0) in m517 CARD01 (chronic pain) looks similar at the K4.1-bit level to m610 i_case_cancer_150 (BRCA1/2 breast cancer, TMU-A synthetic-lethality, also K4.1 (1,1,0)). This does not mean the two conditions share underlying biology. transferability_pct: a same K4.1 bit pattern is necessary, not sufficient, for cross-domain Roll transfer; the underlying X-structures (brain network vs. tumor-cell genome) are categorically distinct. m620 verification ladder (m620 exec dirs) is the path to a quantitative transferability score; no such score is asserted by m625 v0.1.
Falsifying prediction (per W6.5 + Roll EPP-3)
The current evidence pattern, as Roll-formalized in m517 CARD01, would be falsified by either of:
an independent re-cooking of Tu 2020 (or a successor study) under m517's 02_cards_cooked audit pipeline that fails to satisfy the TMU-A K4.1 = (1, 1, 0) declaration; or
an m517 W7-N re-audit on this reading layer that fires any detector (W7-N.1 fabrication / W7-N.4cure-reverse / W7-N.5endorsement).
Both paths are open. Advocates and skeptics are invited to attempt either.
AS-001 (cure lexicon smuggle) — should not occur · audit gate G2.2
AS-006 (digital twin lexicon smuggle) — should not occur · audit gate G2.2
AS-010 (marketing hype smuggle) — should not occur · audit gate G2.2 + W6.2
The audit run on this demo is recorded in `_ops/SELF_AUDIT_LATEST.json` after each `python3 -m 02_audit.audit --self-audit` invocation.
Ferry target (per W6.6)
Primary: m511 clearcancer_circle — page_candidate sync_group_id m625-to-m511-YYYYMMDD-page_candidate-N. m511 is the sovereign for patient-facing public surfaces.
Secondary candidate: m590 research_x1000_ai — claim_pack carrying the m517-anchored TMU-A K4.1 (1,1,0) seed for the knowledge graph.
Frontmatter `ferry_target: m511`. Actual payload generation is deferred to v0.2 (per `05_docs/M625_FERRY_TARGETS.md`).
Honest verdict (per W6.5)
已做:
- Patient-advocate reading layer cooked over m517 CARD01 Tu 2020 attractor.
- K4.1 (1,1,0) declaration referenced by spec section.
- 3 attack-surface entries (AS-001, AS-006, AS-010) linked.
- Cross-domain bit-pattern observation flagged as necessary-but-not-sufficient (per W6.3).
- Falsifying prediction stated with two re-audit paths.
待办:
- Plain-language community version of this reading layer (lower reading level · v0.2).
- Cross-condition synthesis across CARD01 (Tu 2020) + CARD06 (VGAIT) + CARD08 (Wu 2025) for the advocate (deferred to v0.2 after CARD08 cooking · see Demo 006).
- Live advocate walkthrough (deferred to v0.2 owner-led session per m517 "请教 NOT 展示" doctrine).
漏洞:
- The Tu 2020 source is classified as `training_data_recall · TODO_unverified` because m517 itself labels the underlying anchor as drafted-not-signed (m517 self-reported readiness from `01_materials_cache/A5_m517_neuro_anchor.md` is mid-band, not yet signed). Live PubMed verify is open work for m517 v0.4+, not for m625.
- The K4.1 (1,1,0) bit-pattern coincidence with m610 i_case_cancer_150 (BRCA1/2 breast) is observational, not derivational; m620 verification is the path to a substantive cross-domain claim.
- No quantitative effect-size claim is made for chronic-pain intervention; this is by design (m517 W7-N + m625 W6 anti-marketing) and may frustrate advocates expecting numbers.
反驳预案:
- *"Where is the trial number?"* — m625 does not host trial data. m517's outbound `06_m590_neuro_claim_seeds/SEED_PACK_v0.3.md` contains many seeds with NCT/PMID anchors at training_data_recall tier; live verify is m517 v0.4+.
- *"Why no plain English?"* — v0.1 reading layer is researcher-readable; v0.2 will produce a plain-language variant. Filed under "待办".
- *"How is this different from a wellness blog?"* — a wellness blog claims a `cure`. This demo declares `K4.1 = (1,1,0)`, states the falsifying prediction in machine-checkable terms, and exposes the attack surface. A wellness blog does none of these.
- *"Why CARD01 and not the others?"* — CARD01 is the deepest cooked card in m517 v0.3.0 with a clean TMU-A classification. CARD06 (VGAIT) and CARD08 (Wu 2025 cross-condition) are anchored in adjacent demos (Demo 006 covers Wu 2025).