With cell and lobule input counts fixed, 40 released mouse cerebellar cell maps showed no excess of four-zone convergence

We tested whether input from all four cerebellar zones was an excess pattern after connection counts were fixed. We used a public binary map of six lobules and 40 cerebellar nuclear cells from Gruver et al. (2024). Exact counting kept each cell's number of connected lobules and each lobule's number of target cells. Three cells had all four zones. The exact reference mean was 3.3365. Its central 95% range was 1 to 5, and its inclusive upper tail was 0.78895. Assigning lobule IX always to posterior also gave no excess. A secondary score counted co-input in III with IV/V and VIII with IX. The observed score was 14, versus an exact fixed-margin mean of 8.4001. Its sampled upper tail was 0.00038. An added control also fixed total co-input in neighboring lobules. It raised the score mean to 11.7838 and the upper tail to 0.07812, with Monte Carlo standard error 0.00164. The binary map matches both public exports. It omits lobule II, has no linked animal IDs, and differs from the source 41-cell summary. The result covers the released subset and the stated graph references. It does not show random wiring or an absence of cross-zone integration. Exhaustive checks, two samplers, and a clean offline repeat passed. The package preserves code, public source tables, exact counts, all reference draws, figures, provenance, and repeat instructions.

cerebellum · Purkinje cells · functional connectivity · convergence · fixed degree · network motifs · exact counting · structure and function · reproducible research