LIMEN Science: What We Know

UNACCEPTED SCIENCE SNAPSHOT · CONCEPT STUDY · RESEARCH ONLY · NO HUMAN TESTING · NOT AVAILABLE · NO DEVELOPMENT PROGRAM EXISTS

The 19-file package supports continued research, not practical implant development. The exact target—stable multi-year direct nerve recording plus afferent return—remains unproven.

Executive Summary

• Bounded component evidence exists. Human studies have separately demonstrated peripheral-nerve recording and nerve stimulation; small human systems have also combined implanted muscle sensing with nerve stimulation.
• The exact LIMEN target is not established. No accepted evidence in this package shows a chronic, fully implanted human system combining stable direct natural nerve recording with afferent return and proven simultaneous operation.
• The bottleneck is not basic amplification. It is maintaining signal identity, selectivity, calibration, tissue compatibility, artifact separation, and useful advantage over lower-risk external or muscle-based systems for years.
• Next move: research gates, then synthetic software. Close Science QA, define one population and benefit, test the lower-risk comparator case, then authorize R0—not a practical implant.

What Was Researched

The package compared conventional electronic, biohybrid/living-intermediary, and combined biological/electronic interfaces. It inspected human and animal nerve recording/stimulation, muscle-based RPNI systems, engineered axons and iPSC-myocytes, long-duration failures, lower-risk comparators, HAPTIX/HORNET, and six close patent-publication families.

Direct natural ENG means nerve activity, not muscle activity or an evoked response. Simultaneous duplex means usable natural nerve recording during relevant return stimulation with artifact handling explicitly shown; this remains unknown.

Most Decision-Critical Questions Remain Unresolved

Three questions have bounded component or adjacent support. The rest remain open, unknown, hypothetical, or not established. This is why the next phase must close evidence gaps instead of converting the current architecture documents into a physical implant.

All questions is the denominator; bounded support and unresolved partition it (3 + 5 = 8). These are not probabilities, readiness levels, or evidence-quality scores.

What the Evidence Concludes

• Direct nerve recording plus stimulation: demonstrated only as bounded short human components; the task loop used mixed neural plus muscle signals and separate-nerve stimulation.
• Chronic direct nerve recording: open. Longer array residence included major recording-contact loss.
• Long-lived return path: bounded multi-year stimulation-contact evidence exists, but it is not stable natural nerve recording.
• Chronic two-way human systems: adjacent examples use muscle EMG forward and nerve stimulation return—not direct ENG.
• Living intermediary: plausible hypothesis; no unique chronic human two-way advantage is established.
• Simultaneous record/stim operation: unknown for the exact route and interface.
• Advantage over lower-risk systems: unknown. Comparator advantage is a gate, not an assumption.
• Novelty and FTO: not established. The field is already crowded.

Conclusion: LIMEN remains a bounded research platform question, not a chosen implant.

No Architecture Has Earned Selection

• A — conventional electronic: strongest direct component precedent; blocked by chronic direct-ENG stability, artifact recovery, selectivity, and comparator advantage.
• B — biohybrid/living intermediary: real muscle-transducer and animal evidence; blocked by unknown unique contribution, containment, stability, and human two-way benefit. It cannot be assumed to protect native cells.
• C — combined: real adjacent precedents; blocked by failure attribution, coupled drift, maintenance burden, exact timing, and incremental value.

Multi-year stimulation or muscle-sensing records do not establish a repairable, upgradeable multi-year direct-ENG platform. HAPTIX, HORNET, RPNI, living-electrode, and tissue-engineered interface prior art also defeats broad originality language.

What R0 and R1 Are Actually For

• R0: generated synthetic records processed offline by deterministic software. No biological inputs and no actuation.
• R1: generated synthetic signals passed through a non-biological bench harness to a passive load or phantom. No tissue, person, animal, implant, or external actuator.

R0/R1 can test whether the software preserves route labels, rejects ambiguity, quarantines artifacts, separates command from delivery and observation, fails closed, and reproduces the same output from the same fixture. They cannot establish biological detection, safety, efficacy, feasibility, chronic stability, tissue compatibility, encoding, lifetime, reversibility, repairability, or architecture superiority.

What to Do Next

1. Close the Science package gate. Send these exact bytes to fresh Independent Science QA. If QA passes, the Integrator must independently reproduce the hashes and record acceptance. Any changed byte creates a new snapshot.
2. Choose one bounded research problem—not a device. Define one population, one unmet function, one measurable benefit, one exclusion boundary, and one exact forward/return route.
3. Name the comparator that could kill the invasive idea. Compare against optimized external, removable, temporary, muscle-EMG, or usual-care alternatives serving the same function. If the same benefit is available with lower burden, stop the invasive branch.
4. Commission four focused evidence closures. Direct-ENG longevity; record/stimulation artifact recovery and timing; person-specific selectivity/calibration drift; and incremental benefit versus the matched comparator.
5. Only then authorize Backend R0. Build the synthetic conformance system first. R1 follows only after R0 passes its exact accepted contract and Backend, QA, and Integrator gates. Practical implant blueprints, biological protocols, purchasing, assembly, and human/animal work remain blocked.

Verify the Core Evidence

Short human components · longer array follow-up · multi-year stimulation contacts · muscle EMG plus nerve return · biohybrid animal evidence · DARPA HAPTIX

Acceptance State and Limits

The 17Source for 17Source: LIMEN artifact manifestFile: docs/science/limen/ARTIFACT_MANIFEST.md substantive checksums reproduce, but the Science package remains a producer restart candidate without fresh Independent QA or Integrator acceptance. Bounded negative searches are not proof of absence. Nothing here establishes safety, efficacy, feasibility, universality, novelty, patentability, freedom to operate, product selection, or development readiness.