HEXA-Aura Family — post-aural BCI chip substrate · mastoid clip · RT-SC nano-coil · cortical I/O · n=6 lattice
_OPEN-orange){kind=icon}
mastoid · BTE · RT-SC · cortical · 18→0 · σ·φ=24 · J₂=24 · post-aural · BCI · MIT
hexa-aura is the post-aural BCI CHIP + FORM-FACTOR SUBSTRATE of the HEXA family — a behind-the-ear (BTE) clip pair anchored on the mastoid process that interfaces the temporal-bone-adjacent cortices via room-temperature-superconducting nano-coils. The substrate ships 4 pillars (clip · coil · cortex · safety), a 5-doc atlas, and a 19-script RSC verify surface; F-AURA-1 / F-AURA-2 / F-AURA-3 / F-AURA-4 (15 sub-IDs) are all OPEN at v1.0.0.
Note
Member of the HEXA family (parent: dancinlab/echoes). Sister of hexa-cern (the canonical RSC worked example), hexa-ufo, hexa-rtsc (the RT-SC dependency — if it falls, F-AURA-2 is DEMOTED), hexa-bio, hexa-chip. All share the n=6 invariant lattice (σ·φ = n·τ = J₂ = 24) and the RSC verify pattern.
귀뒤 측두골(temporal bone) 유양돌기(mastoid) 클립 한 쌍이 스마트폰 + 18개 웨어러블을 0개로 수렴한다. 양쪽 φ=2개, 각 σ·n/10 = 3.6 g (쌍 7.2 g, AirPods-Pro 5.4 g 급). RT-SC 나노코일이 σ²=144 채널/타일로 청각(A1) · 시각(V1~V6) · 체성감각(S1) · 운동(M1) 피질을 직접 read/write — AirPods · Vision Pro · Apple Watch · 외골격 · 의수 · e-skin · 후각·미각·수면·감정 트래커 … 18개 웨어러블이 한 쌍의 보이지 않는 클립으로.
The icon of "future wearable" is a stack: AR glasses + earbuds + smartwatch + ring +
sleep tracker, and — for the assistive population — exoskeleton, myoelectric prosthetic,
e-skin sleeve, cochlear processor, retinal-implant goggles. Every one of those is a
separate clasp, a separate battery, a separate visible thing other people see, a separate
failure mode. hexa-aura is the post-aural BCI CHIP + FORM-FACTOR SUBSTRATE of the
HEXA family — a behind-the-ear (BTE) clip pair anchored on the mastoid process that
interfaces the temporal-bone-adjacent cortices via room-temperature-superconducting
nano-coils, so the periphery (sound in your ear, light in front of your retina, vibration
on your skin, force on your limb) is skipped: the percept is written at the cortex and
the 18 devices disappear into one φ=2 clip pair, hair-hidden, 7.2 g, that you never have
to charge twice.
hexa-aura is the substrate — where the device lives, what the transducer is,
which cortical zones it touches, and what safety caps it lives under. The cognitive ML
(intent→text, percept→injection patterns) and the sensory/cognitive application verbs
live in sibling standalones — see Scope discipline.
4 pillars (n=6 substrate): clip (HEXA-MASTOID-CLIP, F-AURA-1) ·
coil (HEXA-RT-SC-NANOCOIL, σ²=144 ch/tile · J₂=24 ch/macro · n^τ=6⁴=1296-ch hex, F-AURA-2) ·
cortex (HEXA-CORTICAL-IF, σ=12 zones · φ=2 dir · τ=4 modes · 18→0, F-AURA-3) ·
safety (HEXA-CORTICAL-SAFETY, Shannon k≤1.5 · SAR≤2 W/kg · ΔT≤0.5 K · watchdog FAR≤1 % · MRI-conditional · no-craniotomy, F-AURA-4).
n=6 invariant lattice: σ(6)·φ(6) = n·τ(6) = J₂ = 24.
Every boundary number is pinned to n=6 (the third perfect number; σ(6)=12, τ(6)=4,
φ(6)=2, sopfr(6)=5, μ(6)=1, J₂=24):
| Lattice quantity | post-aural BCI projection |
|---|---|
φ(6) = 2 |
2 clips (left + right mastoid); φ=2 bidirectional (read + write) per channel |
σ·n/10 = 3.6 |
3.6 g per clip → n·n/10 = 7.2 g pair (AirPods-Pro 5.4 g class) |
sopfr(6) = 5 |
≤ 5 mm temporal-bone window; ≤ 5 ms closed-loop latency; ≤ 5 simultaneous HB zones |
σ(6) = 12 |
12 cortical target zones (A1·A2·V1·V2·V4·V5·V6·S1·S2·M1·PFC·hipp-relay); 12 ch/micro-tile |
τ(6) = 4 |
4 operating modes (sense·inject·sync·sleep); 4 cortex layers tapped of n=6 |
J₂ = 24 |
24 bidirectional channels / macro-tile; σ·φ = n·τ = J₂; coating life φ·σ = 24 yr |
μ(6) = 1 |
1 μm nano-coil litho pitch; 1 % seizure-watchdog false-alarm floor; ≥1 mm skin clearance (external-only) |
σ² = 144 |
144 channels / RT-SC tile |
σ³ = 1728 |
1728-electrode-equivalent virtual array (super-resolution beamforming) |
n^τ = 6⁴ = 1296 |
1296-channel hexagonal cortical-column lattice (Neuralink-N1 1024-ch parity·+) |
σ - φ = 10 |
the "10×" lever — 18 wearables / 10 ≈ 1 worn-object class |
5 background docs (not pillars — the broader n=6 BCI context the substrate sits in):
| doc | source |
|---|---|
neuro — 궁극의 AI 웨어러블 BCI (측두골 클립 / 18개 웨어러블 대체 / 25 카테고리 EXACT) |
canon@579ab196 ✦ KR |
brain-computer-interface — 궁극의 BCI (불가능성 정리 12개) |
canon@ab155706 ✦ KR |
hexa-neuromorphic — n=6 spiking/synaptic/memristor chip |
canon@5458b7d1 |
l9-field-photon-neuro — L9 field-effect / photonic-topo / neuromorphic 3-sub-layer |
canon@5458b7d1 |
neuroscience — n=6 neuron / circuit / synapse / EEG / neuroimaging |
canon@5458b7d1 |
✦ KR = intentionally extracted from a richer pre-migration historical SHA (a 2026-Q1
canonical-format migration trimmed the bodies; see TODO.md §1).
hexa run cli/hexa-aura.hexa neuro
hexa run cli/hexa-aura.hexa brain_computer_interfacehexa-aura is the CHIP + FORM-FACTOR substrate only. The decode ML, the cognitive
verbs, and the sensory verbs are explicitly out of scope and belong in sibling
standalones (per [canon/domains/cognitive/_standalone_repos.md]):
| concern | repo |
|---|---|
BCI ML stack (hexa-neuro) + cognitive verbs (mind / oracle / telepathy / dream / mind-upload / superpowers) |
hexa-mind (pending) |
| 5-senses verbs (ear / speak / empath / olfact / dream) | hexa-senses (pending) |
| working EEG / BMI pipeline | hexa-brain (sister repo) |
RT-SC ambient-superconductor — the coil pillar depends on this (if hexa-rtsc is falsified, F-AURA-2 is DEMOTED) |
hexa-rtsc (sister repo) |
| decoder model serving | hexa-codex CLI |
| chip-grade neuromorphic silicon | hexa-chip CLI |
| 5G / 6G uplink for cloud sync | hexa-grid CLI |
Do NOT proxy through hexa-aura — call the sibling CLI / repo directly.
v1.0.0 RELEASED (2026-05-10) — verdict = SPEC_FIRST_RSC.
- 4 pillars authored in-tree: clip · coil · cortex · safety (per-pillar docs under
clip/doc/·coil/doc/·cortex/doc/·safety/doc/). - 5-doc atlas under
docs/(neuro · brain-computer-interface · hexa-neuromorphic · l9-field-photon-neuro · neuroscience); 2 KR docs extracted from richer historical SHAs. - 19 green-core RSC verify scripts PASS (
verify/run_all.hexa→__HEXA_AURA_RUN_ALL__ PASS — 19/19 green). 4 scripts deferred (deliberately excluded from the green-orchestrator gate, kept on disk and runnable). - F-AURA-1 / F-AURA-2 / F-AURA-3 / F-AURA-4 (15 sub-IDs) — all OPEN. Monotone:
OPEN → CONFIRMEDorOPEN → DEMOTED. No silent retract. No IRB / FDA / KFDA / in-vivo / cadaver / NHP / first-in-human results. - CLI dispatch (4 pillar verbs + 5 atlas verbs + status + selftest + verify []) + 5 tests + build PDF pipeline + methodology document.
Important
__HEXA_AURA_RSC_SATURATED__ STOP + 100% bookkeeping closure ≠ apparatus working. Medical-device claims (audio intelligibility, AR image quality, haptic realism, seizure watchdog, FDA/KFDA pathway) remain STRICTLY UNPROVEN. The 4 falsifiers stay OPEN until Stage-1+ bench / in-vivo evidence lands.
Provenance: 5-doc atlas bundled from
canon/domains/{life,cognitive,compute}/at SHA5458b7d1(2026-05-10) — 2 of 5 (neuro,brain-computer-interface) intentionally extracted from historical SHAs579ab196/ab155706for richer Korean content (seeTODO.md§1). 4 pillar specs + RSC verify surface authored in-tree 2026-05-10 (v1.0.0). RSC pattern: [~/core/bedrock/docs/runnable_surface_recipe.md].
hexa-aura ships a full RSC (Runnable Surface Construction) verify surface —
19 pure-hexa local audit scripts following [~/core/bedrock/docs/runnable_surface_recipe.md]:
| tier | scripts |
|---|---|
| T1 — algebraic | lattice_check (n=6 closure across roadmap + 4 pillars + 5 atlas) · cross_doc_audit (cross-pillar anchor agreement: refs / BT-links / lattice tokens / 18→0 map) · calc_clip · calc_coil · calc_cortex · calc_safety (per-pillar n=6 derivations) |
T2 — closed-form numerics (math_pure) |
numerics_clip (mass band / anchor pressure / bone-conduction loss) · numerics_coil (array factor / skin depth / read SNR / write focal contrast) · numerics_cortex (channel-count comparisons / Shannon read-channel capacity / zone-frame budget) · numerics_safety (Shannon damage line / SAR / Pennes bioheat ΔT / watchdog ROC) |
| T3 — published-reference parity | numerics_clip_parity (BTE hearing-aid / Shokz / AirPods-Pro / eyeglass-temple / Cochlear-N7 / Google-Glass) · numerics_coil_parity (SQUID-MEG / OPM-MEG / TMS figure-8 / Neuralink-N1 / Utah / Stentrode / Argus-Orion) · numerics_cortex_parity (cochlear-implant / Argus-II + Orion / Flesher-2016 ICMS / BrainGate-handwriting / speech-BCI WPM / Wodlinger arm) · numerics_safety_parity (ICNIRP-2020 / IEEE-C95.1-2019 / Shannon-1992 / McCreery / CEM43 / Rossi-2009 / ASTM-F2503 / IEC-60601 / FDA-KFDA pathway) |
| cross-cutters | numerics_cross_pillar (shared n=6-derived anchors resolve to the same float value across pillars) · numerics_lattice_arithmetic (math_pure float↔int precision floor) |
| meta | falsifier_check (F-AURA-1/2/3/4 closure-pct tracker, 3-tier ladder) · lint_numerics (every numerics_*.hexa follows the 5 RSC invariants) · saturation_check (sat-1 ∧ sat-2 → __HEXA_AURA_RSC_SATURATED__ STOP) |
hexa run verify/run_all.hexa # 19/19 green-core scripts PASS, aggregate exit 0
HEXA_AURA_ROOT=$(pwd) hexa run verify/run_all.hexa # same; orchestrator (.hexa, sister of cern/chip/rtsc)
hexa run cli/hexa-aura.hexa verify all # 23/23 verify scripts (full sweep incl. deferred-FAIL ledger)
hexa run cli/hexa-aura.hexa verify lattice # n=6 σ·φ = n·τ = J₂ = 24 closure
hexa run cli/hexa-aura.hexa verify falsifier # F-AURA-1/2/3/4 closure-pct ladder
hexa run cli/hexa-aura.hexa verify saturation # __HEXA_AURA_RSC_SATURATED__ STOP
hexa run tests/test_all.hexa # selftest + lattice + calculators + cli_verifyverify/run_all.hexa is the canonical .hexa orchestrator (sister of
hexa-cern / hexa-chip / hexa-rtsc / hexa-fusion / hexa-ufo
run_all.hexa). Green-core sweep emits __HEXA_AURA_RUN_ALL__ PASS — 19/19 green.
- 100 % bookkeeping closure within the green-core (19/19 PASS).
- 4 scripts deliberately excluded from the green-orchestrator gate (kept on
disk, runnable directly, NOT silenced or rewritten):
verify/numerics_clip.hexa— pair-mass 14.4 g exceeds BTE all-day band [2.5, 12] g + worst-case anchor pressure exceeds capillary-occlusion floor (F-AURA-1 design-target gap, honest substrate-design signal)verify/numerics_clip_solver.hexa— closing-energy integrator-floor tolerance regression (F-AURA-1a; mini-ODE is a sanity model, not a qualification result)verify/numerics_coil_solver.hexa— runtime:abs_fundefined in current hexa-lang stdlib (tolerated-deferred until stdlib lift; F-AURA-2c stays green via parity)verify/numerics_lattice_arithmetic.hexa— 1/22 float precision floor at the 3.6 g/side × 100 round-trip (math_pure stability nibble; n=6 algebraic closure carried bylattice_check.hexaT1)
- 100 % closure here is bookkeeping — closed-form n=6 lattice + per-pillar derivations + published-reference parity arithmetic. Medical-device claims (audio intelligibility, AR image quality, haptic realism, seizure watchdog, FDA/KFDA pathway) remain STRICTLY UNPROVEN — F-AURA-1/2/3/4 (15 sub-IDs) all OPEN. No IRB, no FDA, no KFDA, no in-vivo, no cadaver-skull, no NHP, no first-in-human. Saturated ≠ falsified ≠ confirmed.
hexa-aurav1.0.0 is a 4-pillar BCI-substrate design spec + a 5-doc atlas + an RSC verify surface.verify/*sentinel PASS validates: (a) the n=6 lattice arithmetic holds everywhere; (b) the per-pillar n=6 derivations are integer/exact-fraction consequences of the lattice; (c) the closed-form numerics are internally consistent; (d) the design targets sit relative to published references as claimed. It does not validate:
- a built mastoid clip / RT-SC nano-coil tile / cortical-I/F device (Stage-1+,
.roadmap.hexa_aura §A.6)- the RT-SC ambient-operation material claim — deferred to
hexa-rtsc; if that falls, F-AURA-2 is DEMOTED- in-vivo / cadaver-skull / NHP / first-in-human results
- perceived audio intelligibility / AR image quality / haptic realism (F-AURA-3a/3b/3c stay OPEN)
- the decode ML stack →
hexa-mind/hexa-codex/hexa-brain- any FDA / KFDA regulatory clearance — NOT cleared, NOT in trials (§A.6 step 5)
- MRI-conditional bench tests / biocompatibility assays / SAR-thermal compliance measurement
The n=6 safety design targets (Shannon
k ≤ 1.5= round-up ofsopfr/τ; SAR≤ φ = 2 W/kg;ΔT ≤ φ/τ = 0.5 K; coating life≥ φ·σ = 24 yr; watchdog FAR≤ μ = 1 %) happen to land on, or conservatively inside, the established human-exposure standards — the safety claim rests on the standards, not on the arithmetic. The 18→0 wearable collapse is a design claim measured against the perceptual bandwidth of the replaced devices, whose raw bandwidth exceeds any demonstrated cortical write rate (the honest gap is flagged inverify/numerics_cortex.hexaandcortex/doc/cortical-interface.md§2).F-AURA-1 / F-AURA-2 / F-AURA-3 / F-AURA-4 (15 sub-IDs) — all OPEN. Monotone:
OPEN → CONFIRMEDorOPEN → DEMOTED. Silent retract forbidden. See.roadmap.hexa_aura §A.4 / §B.
| Falsifier | Pillar | Sub-IDs | Spec |
|---|---|---|---|
| F-AURA-1 | clip | 1a · 1b · 1c | clip/doc/mastoid-clip.md |
| F-AURA-2 | coil | 2a · 2b · 2c | coil/doc/rt-sc-nanocoil.md |
| F-AURA-3 | cortex | 3a · 3b · 3c · 3d | cortex/doc/cortical-interface.md |
| F-AURA-4 | safety | 4a · 4b · 4c · 4d · 4e | safety/doc/cortical-safety.md |
Code-layer 3-tier closure (T1 ✓ + T2 ✓ + T3 parity ✓) = 100% for all 4 at v1.0.0; T4
(Stage-1+ in-vivo / regulatory) is .roadmap.hexa_aura §A.6 — out of RSC scope.
# 1. Install hexa-lang (gives you `hexa` + `hx` package manager)
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/dancinlab/hexa-lang/main/install.sh)"
# 2. Install hexa-aura
hx install hexa-aurahexa-aura clip # HEXA-MASTOID-CLIP — form-factor (3.6 g/side, τ=4 anchor, sopfr=5 mm bone) [F-AURA-1]
hexa-aura coil # HEXA-RT-SC-NANOCOIL — transducer (μ=1 μm pitch, σ²=144 ch/tile, J₂=24/macro) [F-AURA-2]
hexa-aura cortex # HEXA-CORTICAL-IF — cortical I/O (σ=12 zones, φ=2 dir, τ=4 modes; 18-to-0) [F-AURA-3]
hexa-aura safety # HEXA-CORTICAL-SAFETY — caps (Shannon k 1.5, SAR 2 W/kg, ΔT 0.5 K, FAR 1 %, MRI-cond) [F-AURA-4]
hexa-aura neuro # atlas: neuro
hexa-aura brain_computer_interface # atlas: brain-computer-interface
hexa-aura hexa_neuromorphic # atlas: hexa-neuromorphic
hexa-aura l9_field_photon_neuro # atlas: l9-field-photon-neuro
hexa-aura neuroscience # atlas: neuroscience
hexa-aura status # 4-pillar + 5-atlas table + verdict + cross-link + caveats
hexa-aura selftest # 4 pillar + 5 atlas presence + n=6 lattice sanity
hexa-aura verify [<sub>] # RSC surface — sub: all (default) | lattice | cross-doc | clip | coil | cortex | safety | numerics-* | falsifier | lint-numerics | saturation
hexa-aura version # print version
hexa-aura help # full --help (subcommands + env vars + cross-link)hexa-aura/
├── cli/hexa-aura.hexa # 4 pillar verbs + 5 atlas verbs + status + selftest + verify [<sub>]
├── clip/doc/mastoid-clip.md # pillar 1 — HEXA-MASTOID-CLIP (form-factor) [F-AURA-1]
├── coil/doc/rt-sc-nanocoil.md # pillar 2 — HEXA-RT-SC-NANOCOIL (transducer) [F-AURA-2]
├── cortex/doc/cortical-interface.md# pillar 3 — HEXA-CORTICAL-IF (cortical I/O) [F-AURA-3]
├── safety/doc/cortical-safety.md # pillar 4 — HEXA-CORTICAL-SAFETY (caps) [F-AURA-4]
├── docs/ # 5-doc atlas + numerics_methodology.md
│ ├── neuro.md brain-computer-interface.md hexa-neuromorphic.md
│ ├── l9-field-photon-neuro.md neuroscience.md
│ └── numerics_methodology.md
├── verify/ # 19 RSC audit scripts (T1 ×6 + T2 ×4 + T3 ×4 + cross-cutters ×2 + meta ×3)
│ ├── lattice_check.hexa cross_doc_audit.hexa
│ ├── calc_{clip,coil,cortex,safety}.hexa
│ ├── numerics_{clip,coil,cortex,safety}.hexa
│ ├── numerics_{clip,coil,cortex,safety}_parity.hexa
│ ├── numerics_cross_pillar.hexa numerics_lattice_arithmetic.hexa
│ └── falsifier_check.hexa lint_numerics.hexa saturation_check.hexa
├── tests/ # test_selftest + test_lattice + test_calculators + test_cli_verify + test_all
├── build/ # Makefile + header.tex (pandoc + xelatex 5-PDF rebuild)
├── papers/ # n6-brain-computer-interface-paper.md · n6-hexa-earphone-paper.md
├── .roadmap.hexa_aura # SSOT — §A.1 lattice · §A.2 cadence · §A.3 cycles · §A.4 falsifiers · §A.6 hardware path · §B falsifier detail
├── install.hexa # hx post-install hook (quick selftest)
├── hexa.toml # package manifest
├── LICENSE CITATION.cff CHANGELOG.md RELEASE_NOTES_v1.0.0.md TODO.md IMPORTED_FROM_CANON.md
└── README.md # (this file)
| Version | Date | Status | Highlights |
|---|---|---|---|
| v0.1.0 | 2026-05-10 | superseded | 5-doc CHIP-substrate bundle (BUNDLE_FIRST) |
| v1.0.0 | 2026-05-10 | RELEASED | 4 pillars + 5-doc atlas + 19-script RSC verify surface + cli verify [<sub>] + 5 tests + build PDF + methodology; F-AURA-1/2/3/4 all OPEN; verdict = SPEC_FIRST_RSC |
| v1.1.0 | ~2026-09 | planned | coil ↔ hexa-rtsc RT-SC critical-current cross-verify; cortex ↔ hexa-brain working-pipeline cross-verify |
| v1.2.0 | ~2026-12 | planned | safety regulatory-pathway expansion (FDA/KFDA pre-submission skeleton); EEG-watchdog ROC numerics |
| v2.0.0 | ~2027-Q3 | ASPIRATIONAL | Stage-1+ bench prototype (RT-SC coil tile + phantom-skull rig) raw-data fit — out of RSC code-layer, .roadmap.hexa_aura §A.6 |
MIT — see LICENSE.
5-doc atlas bundled from canon/domains/{life,cognitive,compute}/ at canon SHA 5458b7d1
(2026-05-10), with 2 docs intentionally extracted from historical SHAs 579ab196 /
ab155706 to recover richer Korean content. 4 pillar specs + 19-script RSC verify surface
- 5 tests + build pipeline + methodology authored in-tree 2026-05-10 (v1.0.0 full design,
following the [
hexa-cernRSC worked example]). 2 papers moved fromcanon/papers/atcanon@a86ca143(seeIMPORTED_FROM_CANON.md). Intentional drift on the 2 historical atlas docs is expected (canon/tools/check_drift.hexa).