Cell, distance and route controls reduced an inhibitory path weight difference from 10.15% to -0.10% in one mouse V1 scan

We tested whether a structural inhibitory path difference between activity assemblies depends on route count, relay cells or cell-level differences. We used the archived activity relation for 61 excitatory cells in one mouse V1 scan and rebuilt the paths from 137,706 source synapses. All 3,657 positive inhibitory path weights matched the source tables. The disjoint-to-shared geometric mean ratio was 1.10073 for total path products, 0.99930 for route count and 1.10150 for mean product per route. Separate source and target cell terms reduced the mean product difference from 10.15% to 1.77%. Soma distance and route count controls reduced it to -0.10%. A whole-cell label test within 50 micrometre depth groups gave a two-sided tail fraction of 0.957 from 4,999 draws. The raw total-product difference stayed positive after each of 215 active relay removals, with a range of 8.21% to 12.47%. An added diagnostic gave ordinary independent-pair rank-test rejection in 61.67% of unrestricted random-label draws at a 5% threshold. This diagnostic concerns one fixed graph. The controlled result does not show that biological inhibition is absent. It separates pair selectivity from differences carried by cells under the stated model. Twenty archived endpoint cells were absent from the supplied registration table, so the test uses the archived pair relation and does not independently confirm the activity join. The package includes code, compact inputs, source hashes, results, figures and repeat instructions.

connectome · activity assemblies · inhibitory paths · V1 Deep Dive · network dependence · source and target controls · reproducibility