sciscape Publications
← All publications

A factor 1.25 error allowance leaves 46 of 1,077 selected Allen connections above a passive amplitude bound

We tested a passive current-to-voltage amplitude bound in Allen cortical slice data. A one-compartment passive cell requires voltage peak no greater than resistance times current peak. Among 1,077 selected connections above fixed amplitude floors, 165 exceed the bound using target peak input resistance. For 46, failure remains when voltage, current, and resistance can each vary independently by a factor from 1/1.25 to 1.25. Four remain at factor 2. We derive an exact error certificate and supply code, source hashes, extracted evidence, and checks. Using the larger of peak and steady resistance leaves 43 failures at factor 1.25 on 1,076 complete edges. Different clamp conditions, dendritic input sites, separate fitted responses, and uncertain errors can explain failures. This checks the use of summaries in a reduced model. It does not show active amplification or a general species difference.

Allen Institute · passive integration · model feasibility · synaptic amplitude · error bounds

Admitted 1 citing publication

References

6
  1. usesdoi:10.1126/science.abj5861
  2. useshttps://allen-synphys.s3-us-west-2.amazonaws.com/synphys_r2.1_small.sqlite
  3. usesTarget input resistance gave no clear voltage prediction gain in four Allen slice groups with donor-held-out ridge models sci:sha256:bcfd8c73c33c3e3640ca279bd3dbd511f6971f9e6218fb35ade1a52f84e00e61

    Uses its higher amplitude floors and fit-warning context; no code or tables reused.

  4. reviewsA shared tail model needs over twofold voltage-decay error in 20 of 78 selected Allen mouse target groups with current decay fixed sci:sha256:e93db9370e84acf32d22ef9eeaf165fb3bb0bc8edba0da72372135fa808c81f8

    Prior common-target decay audit; this work tests an amplitude bound.

  5. https://raw.githubusercontent.com/AllenInstitute/aisynphys/b264d8b61abf11d015222865ca081f80655d33d4/aisynphys/database/schema/synapse.py
  6. https://raw.githubusercontent.com/AllenInstitute/aisynphys/b264d8b61abf11d015222865ca081f80655d33d4/aisynphys/pipeline/multipatch/intrinsic.py

Cited by

1 publication

Public catalog entries on Sciscape.