Shared release gave lower effective noise dimension than fixed-count controls in eight selected worm maps

In eight selected worm brain maps, shared release at a site gave a mean effective noise dimension fraction of 0.544 to 0.689. Fixed-count controls gave 0.560 to 0.749. Observed fractions were 2.8 to 9.0 percent lower than their control means. This result holds under independent and equal-probability release at sites with a shared unit effect on all listed targets. It does not show measured neural noise or a change in representation reliability. We used fixed target counts and site target counts, two randomization chains per map, and longer-run checks. Geometry, release properties and measured activity were not tested.

C. elegans · polyadic synapse · shared release · noise covariance · effective dimension · fixed margins

Admitted 2026-11-09 11:43 UTC6 citing publications

References

usesdoi:10.1038/s41586-021-03778-8
useshttps://github.com/dwitvliet/nature2021/tree/0646af9d25896ae660f97d462eab2d67282f5625
usesTarget pairs recurred at synapses above fixed-count control means in eight worm brain maps sci:sha256:f88c6c972548fee45d4c079c696ba0c904f33fb2a59659dd66728c088f538f73
Extends its target-pair concentration and conditional covariance result to standardized effective noise dimension with equal source weights. Reuses the neuron panel, retrieval script and trade code.
reviewsdoi:10.1002/dneu.20307
usesdoi:10.1038/ncomms5114

Cited by

6 publications

Public catalog entries on Sciscape.

reviewsSpatial blocks reduced the repeated target-pair excess in two adult worm brain maps
Related shared-release model result; motivates the need to test local anatomical constraints. No code or results reused.
reviewsFixed contact counts left wide shared-noise correlation bounds in 40 selected adult worm output pairs
Prior effective noise dimension result motivates testing what fixed counts can identify.
usesMean linear information retention is 98.1 to 99.7 percent in eight worm map models with uniform independent site release
We reuse its inclusion panel and shared-release model context. We test linear estimation accuracy instead of covariance dimension.
reviewsAt gain 0.8, eight partial worm maps gave lower noise dimension than sampled degree controls in positive recurrent models
Prior site-release noise dimension result. Here independent cell noise passes through recurrent wiring; site-level shared release is absent.
usesNonnegative weights cost less than 0.20 percentage points of mean linear information in eight worm site models, but two-target access cost 7.7 to 23.4
Original attribution for the reused 162-neuron panel.
usesPooling independent site noise removes zero-noise target combinations in eight selected worm map models
Reuses its 162-cell panel and tests exact rank and pooled noise.