Common inhibitory input showed no clear increase within activity assemblies in 61 cells from one mouse V1 scan

We tested whether excitatory cells in the same measured activity assembly share more inhibitory input cells. We used one mouse primary visual cortex scan with matched optical activity and electron microscopy. The archived relation covered 61 cells and 1,830 distinct pairs. A binary input cosine had mean 0.45957 for shared pairs and 0.46045 for disjoint pairs. A model with cell terms, soma distance and depth difference gave a shared-minus-disjoint contrast of -0.00378. Complete neuron-label moves within 50 micrometre depth groups gave a two-sided tail fraction of 0.2964 from 9,999 draws. Count weights, volume weights, a two-contact threshold and single-cell removal checks kept the negative direction. An added check rebuilt optical memberships for 39 cells with an unambiguous registration. All 741 pair labels agreed with the archive. That subset gave a contrast of -0.0000062 and tail fraction 0.9993. The score equals the correlation of linear inhibitory drives only under independent source noise and equal connection effects. Neural currents and noise correlations were not measured. The result limits a simple common input rule in this selected scan. It does not establish zero biological selectivity. The package preserves source data, provenance, code, all scores and label draws, figures, a local plan, source checks and repeat instructions.

functional connectomics · activity assemblies · inhibitory input · common input · mouse visual cortex · node label reference