STRC ABR transfer function model

QUARANTINED 2026-08-20 — this model may not be used to state a hearing outcome. NOT-MODELABLE: ABR transfer function calibration literature is absent from corpus (Bredberg 1968, Schuknecht & Gacek 1993, "Sun 2024 OTOF Cohort 2" all missing or misattributed); scripts BROKEN; scalar cannot yield audiogram..

Audit STRC Cure-Route Audit 2026-08-20 found:

  • The calibration literature is absent from the corpus. Bredberg 1968 and Schuknecht & Gacek 1993 have no paper notes in sources/lit/; “Sun 2024 OTOF Cohort 2” appears to be a misattribution of Lv 2024. Under AGENTS.md §0c the transfer function is therefore unsourced — this is the phantom-citation failure class that triggered the literature-first rule.
  • The implementing scripts do not run. models/mrna_lnp_audiogram_rescue.py and models/strategy_b_audiogram_rescue.py both exit 1 on stale pre-vault-split roots; both are marked BROKEN 2026-08-20 in STRC Computational Scripts Inventory.
  • A scalar cannot yield an audiogram. The model maps a single functional-OHC fraction to one threshold, but the cochlea is tonotopic and an audiogram is a curve over frequency; a dead region at one place is not compensated by a healthy region elsewhere, and round-window delivery is basal-to-apical non-uniform. This holds even if the calibration were perfect.

No dB or audiological-category claim derived from this model may be quoted until the provenance-and-runtime audit named in that review closes. If the inputs cannot be grounded, the correct result is NOT-MODELABLE, not a recalibrated band.

Goal. Quantitative bridge: “X% OHC rescue” → “Y dB ABR shift”. Needed to decide dose/coverage in delivery hypotheses.

Core inputs.

  • Cochlear amplifier gain per OHC column (~35-50 dB total, distributed)
  • OHC density × rescue fraction → active OHC count
  • Saturation of amplifier (diminishing returns at high rescue)

Canonical reference. Based on fits to 2021 Holt (mouse) + 2026 Iranfar + 2025 DB-OTO CHORD clinical data. Script: ~/STRC/models/abr_transfer_model.py (see STRC Computational Scripts Inventory).

usage

Invoked in Phase 4 PKPD audits across h04 / h06 / h09 / h10 to translate delivery metrics into predicted clinical endpoint.

Connections