Sparse searches matched the LICONN pairing rule for 151,267 mouse CA1 puncta with a 30-voxel limit
We tested a sparse form of the released LICONN pairing rule. The mouse CA1 source had 79,291 pre puncta and 71,976 post puncta. At a limit of 30 raw voxels, the method gave 68,955 links. An independent check of all 5,707,049,016 distances gave the same nearest choices and map. Three new processes used at most 98.43 MiB of memory. The source cost and map arrays alone would need 70.26 GiB. This storage value is calculated. In a separate test of 242 annotated links, the sparse and source rules had a link F1 score of 99.38%. They retained all six annotated common-input pairs and added two extra pairs. A separate one-to-one assignment had a link F1 score of 98.11% but retained none of those six pairs. With independent unit-variance pre signals and unit linear links, common pre sites set off-diagonal post covariance. Thus, a high link score did not establish retention of common input in this test. The original synapse-map rule supports sharing. The contribution is a checked tool and a conditional signal audit. It does not validate the larger biological map or measure neural activity. The package has new code, aggregate results, figures, source hashes, and full repeat steps. Restricted source files stay external.
connectomics · LICONN · synapse matching · nearest neighbors · shared input · covariance · sparse computation · reproducible research
References
5- usesdoi:10.1038/s41586-025-08985-1
We test the released LICONN pairing rule and unchanged coordinate and partner annotations from its pinned public repository.
- doi:10.15479/AT:ISTA:18697
Related original data record and source reuse terms. The exact input bytes come from the pinned Git repository and remain external.
- usesdoi:10.1038/s41592-019-0686-2
The new tool uses SciPy spatial trees, distances, sparse matrices, and a small comparison assignment.
- Shared sites changed variance gain limits by at most 0.0031 in eight nonnegative C. elegans count models with input normalization and independent 50% site success per step sci:sha256:a09e778d760bf30ace40ad39499d291796fb84ead3b40842c852e8432cdb7e62
Prior conditional models show why site sharing matters for covariance. We read this work and reuse no data or code.
- With equal rates and area weights, independent site states gave 34% more summed target variance than independent contacts in olfactory receptor inputs from one fly larva sci:sha256:91917ec12bc0621c63ddd0312989ca8946b930b736f85d3bf9e9591127756e51
Prior conditional contact and site variance analysis. This study instead checks the construction and representation of imaging-derived partners.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.