Selected MICrONS pairs leave a response reliability bias in connection direction uncertain

We tested whether one-way synaptic connections point toward cells with more reliable movie responses in one MICrONS mouse. Both cells had to have measured contact opportunity toward the other cell. The complete sample had 128 pairs from 96 cells. A directional logistic model adjusted for two opportunity-length ratios, layer, and area group. The reliability odds ratio was 1.662 per standardized unit, with a nominal 95 percent cell bootstrap interval of 0.265 to 25.342. Of the connections, 61 pointed toward the cell with higher raw CCmax. The mean target-minus-source CCmax was -0.00701, with interval -0.04528 to 0.03294. The data leave the direction effect uncertain. They do not show a zero effect. Selected cells, one mouse, possible reconstruction error, measurement error, unmeasured cell traits, and unstable bootstrap fits limit inference. Code, source hashes, derived pairs, uncertainty draws, a figure, and retrieval steps support reproduction.

MICrONS · connection direction · response reliability · paired test · null result · connectomics