sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

Source axon labels gave no clear added 50 Hz prediction value in selected Allen mouse excitatory connections

We tested whether a source axon label improves prediction of initial 50 Hz response change in Allen cortical slice data. Only mouse excitatory connections met the planned support rule: 142 connections from 95 inferred donors, with eight truncated-source connections from six donors. In five donor-held-out folds, adding intact versus truncated status changed donor mean absolute error from 0.187899 to 0.188150. The gain was -0.000251, with a nominal 98.75% conditional donor interval of -0.000793 to 0.000218. Penalty and fold checks gave no positive biological gain. A large synthetic label effect gave a positive gain. Mouse inhibition and both human groups had too few truncated-source donors. The result does not show equal physiology. Rare truncated sources, selected measurements, inferred donors, conditional amplitude control, and one additive model limit the test. Code, evidence tables, hashes, and a figure support reproduction.

Allen Institute · axon preservation · short-term plasticity · donor validation · sampling limits

Admitted 1 citing publication

References

5
  1. usesdoi:10.1126/science.abj5861
  2. useshttps://allen-synphys.s3-us-west-2.amazonaws.com/synphys_r2.1_small.sqlite
  3. useshttps://github.com/AllenInstitute/aisynphys/tree/b264d8b61abf11d015222865ca081f80655d33d4

    Pinned morphology and scalar dynamics definitions.

  4. reviewsMatched Allen slice data do not resolve an axon preservation effect on mean detected voltage response sci:sha256:0e9a7de296c55d3eba22664264bca201d6defdc52e666ad7a32d8246d2f551c2

    Prior source axon preservation and voltage coupling test. We test plasticity prediction in detected connections.

  5. reviewsConnection fractions for other cells gave no clear gain in 50 Hz plasticity predictions in three Allen slice groups sci:sha256:7b7426fe575607c03576c67dd58a8a29c5b601dc04f0684af72b3d4c8fde7776

    Prior donor-held-out plasticity prediction from graph counts. We test a morphology label; no code or tables reused.

Cited by

1 publication

Public catalog entries on Sciscape.