limen Step 5 Founder Decision: No Selection

CONCEPT STUDY RESEARCH ONLY NO HUMAN TESTING NOT AVAILABLE NO DEVELOPMENT PROGRAM EXISTS

STEP 5Source for 5Source: Scope-limited Science Integrator acceptanceFile: docs/review/evidence/limen-step5-science-r2-2026-08-23/02-integrator-acceptance.md RESEARCH RESULT: NO-SELECTION.

No target, architecture, device, product, blueprint, simulation, build, test, or development program is selected or authorized. The project remains REVISION; release is unauthorized.

Executive Summary

• Keep one question open; choose nothing. S5R2-T01 is the closest deferred tuple: vision-occluded contact/release object manipulation for adults with acquired unilateral transradial limb loss. It is not a selected target.
• The comparison is unresolved on both sides. The direct-nerve candidate outcome is UNKNOWN; the removable surface-muscle comparator outcome is UNKNOWN; the exact complete-match search is OPEN. The deterministic result is HOLD_BOTH_UNKNOWN (accepted Science decisionPrivate repository source).
• Two narrow component records do not close the comparison. PMID 33725672 removes a uniqueness claim only for its N=1Source for 1Source: Accepted limen Step 5 Science target recommendationFile: docs/science/limen/STEP5_TARGET_SELECTION_R2/TARGET_RECOMMENDATION.md, three-recording-day forward endpoint. PMID 26584497 supplies an N=5Source for 5Source: Accepted limen Step 5 Science target recommendationFile: docs/science/limen/STEP5_TARGET_SELECTION_R2/TARGET_RECOMMENDATION.md, one-month external-feedback component, not a complete-route head-to-head result.
• Next work is evidence work. Engineering, R0, physical work, and practical testing remain blocked.

Words Used Here

ENG is electrical activity recorded from a nerve; here it means direct natural peripheral nerve activity, not muscle activity or an evoked response. EMG is electrical activity recorded from muscle; surface EMG is removable sensing through the skin. The forward path turns a recorded signal into a computer command. The return path conveys information back to the person; afferent peripheral-nerve stimulation is the deferred candidate return. A matched comparator serves the same population, function, endpoint, schedule, and duration with lower burden. A complete-route comparison tests both directions together. HOLD_BOTH_UNKNOWN means neither candidate nor comparator has an eligible complete result. A/B/C mean conventional electronic, biohybrid/living-intermediary, and combined biological/electronic architecture families; all remain unselected.

Route Under Comparison

Deferred candidate route Natural peripheral ENG → intended-state computer command → vision-occluded contact/release task ← contact/release information through afferent peripheral-nerve stimulation

Matched lower-risk comparator Removable surface EMG → intended-state computer command → the same vision-occluded contact/release task ← external vibrotactile or electrocutaneous contact/release feedback

This is a research comparison frame, not an architecture, interface, procedure, or build specification. [HYPOTHESIS · S5R2-CR-026]

Why the Rule Returns Hold

PMID 33725672 reports one transradial participant, simultaneous direct-neural and surface-EMG acquisition, and similar force/velocity decoding across three recording days. Direct-neural uniqueness therefore FAILS only at that narrow forward endpoint; the source has no return path, chronic adequate-duration result, or complete-route advantage. [FACT/FAIL · S5R2-CR-004/005]

PMID 26584497 reports five transradial participants using myoelectric control with external vibrotactile contact/release feedback at home for one month. It is a comparator component, not an eligible head-to-head complete-route comparison. [FACT · S5R2-CR-020]

No eligible same-population complete comparison was located under the frozen rule. That bounded negative search remains OPEN, not proof of absence. Candidate and comparator outcomes remain UNKNOWN; the accepted inference is HOLD_BOTH_UNKNOWN, so the result is NO-SELECTION. [OPEN/UNKNOWN/INFERENCE · S5R2-CR-019/027/028]

Five Candidate Questions, Zero Selections

• T01 — upper-limb contact/release manipulation: DEFER; closest tuple; HOLD_BOTH_UNKNOWN.
• T02 — computer access for severe motor impairment: DEFER; subgroup, route, and comparator remain open.
• T03 — bladder-state monitoring and modulation: HOLD.
• T04 — treatment-refractory post-amputation pain: HOLD.
• T05 — nerve-recovery monitoring and rehabilitation feedback: HOLD.

For T03–T05, HOLD is not evidence FAIL. Their eligible-comparison searches are OPEN, and candidate outcomes are UNKNOWN.

Product and Architecture Standing

The accepted Product snapshot contributes a neutral functional frame: person-specific information flow, comparator logic, lifecycle questions, and maturity gates. It does not supply a target. Architecture families A, B, and C remain unselected. No route identity may be substituted for another, and no functional frame may be converted into a blueprint or development claim. (Product acceptancePrivate repository source)

Legal, Ethics, and Regulatory Standing

The accepted Legal, Regulatory and Ethics snapshot remains an independent gate. It admits no target or architecture and supplies no legal advice, ethics approval, regulatory classification, pathway decision, patentability analysis, or FTO opinion. Intended use, comparator burden, consent, privacy, cybersecurity, reversibility, access, dual use, classification facts, and responsible entities remain unresolved or blocked. (Legal acceptancePrivate repository source)

Dependency-Ordered Next Work

1. Independently QA this report package and reproduce its exact bytes.
2. If QA passes, obtain scope-limited Integrator acceptance of this report only.
3. Keep the T01 population, both routes, endpoint, exclusions, and prospective evidence rule frozen while closing the exact complete-match search and unresolved source fields.
4. Admit only eligible same-population, same-endpoint, same-schedule, adequate-duration evidence; preserve denominators, route identity, adverse events, revisions, interruptions, category changes, and recalibration burden.
5. Reapply the deterministic truth table. If both outcomes remain unknown, keep NO-SELECTION.
6. Only a changed, accepted evidence state may trigger fresh Science, Product, and independent Legal/Ethics/Regulatory review.
7. Synthetic-only R0 consideration remains blocked until a target is selected through those gates and separately authorized.

These are REQUIREMENT statements for research governance, not evidence that selection or development is possible.

Status and Limits

PRODUCER CANDIDATE — PENDING INDEPENDENT QA AND INTEGRATOR ACCEPTANCE

This report imports only the accepted Product, Science, and Legal/Ethics/Regulatory snapshots at their recorded scopes. Challenger records are advisory only. The mixed-origin historical candidate package is excluded from authority. Nothing here establishes feasibility, safety, efficacy, clinical validation, novelty, patentability, FTO, universality, approval, availability, readiness, target selection, architecture selection, or development authority.