With chemical cell degrees fixed, shared input counts could vary in only two of 91 mouse V1 Pvalb maps with electrical pairs

We tested whether electrical-positive Pvalb pairs share chemical partners beyond each cell's chemical connection counts. We used Allen Synaptic Physiology r2.1 small. The primary set had 1,471 assessed pairs from 575 mouse V1 recordings. Of these, 100 pairs in 91 recordings had a positive electrical call. We held the electrical calls, assessed directions, and each cell's chemical input and output degrees fixed. Exact graph counts left one chemical graph in 88 of the 91 recordings. Shared input counts could vary in two recordings, and shared output counts could vary in none. The total shared input count was 8. Its exact reference had counts 8, 9 and 10 with probabilities 0.25, 0.50 and 0.25. Even the largest possible count had upper-tail p=0.25. The output count was fixed at 5. Thus, the specified count tests could not detect enrichment at the 5% level for any graph in this reference set. Keeping each cell's degrees fixed within distance bins removed all graph variation. Stricter call and spike checks, and a wider animal rule, left the same test limit. We prove that three or fewer cells cannot change either shared partner count at fixed degrees, and give a four-cell example where input counts can change. Under independent sources with unit variance and equal inhibitory weights, the shared input count equals summed chemical input covariance at electrical pairs. This is a model interpretation, not a measured covariance result. The study gives a limit on these conditional tests in local slice maps. It does not establish absence of shared partner organization in cortex. The package preserves an exact audit tool, aggregate results, figures, source hashes, independent enumeration checks and repeat instructions. Original Allen data remain external.

electrical coupling · Pvalb · mouse visual cortex · chemical synapses · shared input · shared output · graph randomization · fixed degrees · multipatch recordings · study design