Local trial changes gave no clear positive covariance in 118 selected Allen mouse V1 output pairs

We tested whether two outputs from one neuron retain positive covariance after a local time control. We used public Allen mouse V1 slice response fits. The revised exploratory sample had 118 target pairs, 196 detected connections, 91 source neurons, and 70 inferred mice. We paired nearby complete 50 Hz trains within each recovery protocol. Mean normalized covariance was 0.06124 for original train sums and 0.00791 for local trial differences. Nominal 95 percent mouse bootstrap intervals were [0.00990, 0.11254] and [-0.05158, 0.06762]. The paired decrease was 0.05333, with interval [0.01011, 0.09659]. All combined local-change sensitivity intervals included zero. The decrease was uncertain after some controls and did not occur for first pulses alone. An initial trial-matching rule retained only two mice; its result and the later revision are preserved. Independent scalar calculations and graph matching checks passed. These selected voltage fits do not establish independent release, a noise mechanism, or a rule across brain regions and species. We supply source hashes, code, derived evidence, uncertainty draws, and retrieval instructions.

Allen Institute · divergent connectivity · trial covariance · synaptic variability · protocol matching · time control