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

Cleft size weights reduced model orientation prediction error by 1.36% after count and position controls in 3,461 selected neurons from one MICrONS mouse

We tested whether synaptic cleft size adds visual orientation information after input count and position controls. We scanned all 337,312,429 rows of the MICrONS v1507 filtered synapse table. The fixed sample retained 11,615,820 inputs to 3,461 selected neurons, including inputs from sources without recorded nuclei. The size model used square roots of summed cleft volume fractions by source. Both models had input count profiles, spatial contact features, target labels and matched penalty searches. Five spatial outer folds and three inner folds gave mean squared errors of 0.876330 for the count control and 0.864406 for the size model. This was a 1.36% reduction, with a conditional 95% cube resample range of 1.06% to 1.66%. The gain was 1.32% in a second spatial split, 1.34% with imaging sessions excluded and 1.38% on a wider parameter grid. Recorded size placement gave a higher adjusted score than all 199 local size shuffles at 25 micrometres. Those shuffles kept every connection count and each target's local size distribution fixed. Two other scales gave the same direction. Raw orientation scores did not exceed the reference ranges. These results concern a fixed statistical model and estimated orientation in one mouse. They do not prove a causal role for size or transfer to another animal. The package preserves code, exact compact features, all predictions, all reference scores, figures, source hashes, raw-data checks and repeat instructions.

connectome · MICrONS · synaptic cleft size · orientation · structure and function · shared inputs · spatial controls · prediction

Admitted 2 citing publications

References

10
  1. usesShared input identities reduced estimated orientation prediction error by 2.23% after position controls in a fixed test of 3,461 neurons from one MICrONS mouse sci:sha256:73d06e1c0a8fe6bdba8191b8397acb3c264052279ac8cb916e92afb37ae8b7a3

    We reuse its exact count and geometry features, fixed target records, kernel helpers and extraction code. We add cleft size profiles and local size assignment tests.

  2. usesInput segments without a nucleus reduced orientation prediction error in fixed tests on 3,461 MICrONS neurons sci:sha256:294fd60eacb9f05af75fd1a19e307678b2e734a906199bc08d0b9803ba295fcb

    We reuse its fixed 3,461-neuron cohort and complete eligible input count graph, including sources without nuclei.

  3. supportsIn MICrONS mouse visual cortex, connected excitatory neurons had more similar preferred orientations mainly when their synapses were large (14,412 neurons, digital-twin tuning) sci:sha256:07b8a9b83eea2b7443dc99d29805440aefc0d23414bb9c8959decc534824bd92

    Our held-out-neuron test adds prediction evidence consistent with its size-dependent orientation association. The selection and estimand differ.

  4. In one MICrONS mouse, repeated connections had 2.7–12.6% lower log size variance than two local shuffles with 20 micrometre cubes sci:sha256:f1934d2ea0fec536f3150e85955931e6fb255bbbe10f0e06d9f36de7e2bc5245

    This prior study tested size variance within repeated connections. Our test uses summed sizes for input-profile prediction.

  5. usesdoi:10.1038/s41586-025-08790-w

    The MICrONS minnie65 v1507 public static exports supply all structural and target annotation data.

  6. doi:10.1038/s41586-025-08840-3

    This work established relations between MICrONS connection structure, cleft size and functional similarity.

  7. doi:10.1038/nature14182

    Paired-recording evidence links excitatory synaptic strength to visual response similarity.

  8. usesdoi:10.1038/s41586-025-08829-y

    The fixed exported preferred orientations come from the digital twin response model described here.

  9. doi:10.1038/s41586-026-10457-z

    This correction clarifies the response-model architecture. We use exported scalar properties and do not retrain the model.

  10. usesdoi:10.1038/s41586-024-07765-7

    Cell annotations used in the fixed MICrONS cohort derive from the released resource and contributing annotation work.

Cited by

2 publications

Public catalog entries on Sciscape.