Four partial fly Kenyon cell maps had fewer input pairs from one glomerulus than random maps with fixed cell and channel totals

Repeated input from one olfactory channel can let a Kenyon cell reach a threshold without input from another channel. We tested the four partial Figure 1d input maps released by Ellis and colleagues in 2024. Each map has 200 cell rows and 49 glomerular channels. The maps contain 5, 12, 3, and 5 pairs of inputs from the same channel to the same cell. Uniform random allocation with the recorded cell and channel totals gives exact means of 27.87, 29.47, 35.62, and 21.02. The deficit remains after channel totals are fixed within groups of cells with the same mapped input count. All four lower-tail Monte Carlo values are at most 0.00030 after four-test Holm correction under that reference. The planned excess prediction failed. In an equal-input model with threshold two, the recorded maps give 75–91% fewer single-channel responses than the main reference mean. A descriptive readout added after the pair results gives 14–23% more responses that require two channels together. We provide exact reference moments, all random draws, source audits, figures, and an offline repeat. The maps are incomplete. Animal IDs and cell type IDs are absent. Dye filling can be selective. Article and matrix totals differ. These limits prevent a claim of biological input avoidance or measured learning effects.

Kenyon cell · olfactory circuit · input diversity · partial input map · conditional randomization · structure and function