Input resistance gave an uncertain 29% error reduction for electrical coupling ratios in 57 mouse V1 cell pairs

We tested whether measured input resistance can predict unequal electrical voltage transfer in mouse cortex. The main sample had 57 cell pairs, 55 sessions and 53 inferred donor groups from Allen Synaptic Physiology r2.1. Both electrical directions were detected. A fixed resistance ratio predicted the forward to reverse coupling ratio. No parameters were fitted. Mean squared log ratio error fell from 0.5921 to 0.4197, a reduction of 29.1%. The central 95% donor resample range ran from a 7.1% increase to a 52.9% reduction. Thus, the test did not establish a reliable gain. The broader 106-pair mouse sample gave a 3.6% reduction, also with a range across zero. A separate graph audit retained 129 positive reciprocal pairs in 124 sessions. Only one graph had a loop. The other 123 graphs can fit any positive directional ratios with free cell scales, so they cannot test that necessary condition for reciprocity. The one triangle had a forward to reverse product ratio of 0.995123. This fixed-number result has no measurement error estimate. Input resistance and coupling use different pulse summaries, and the small release lacks source recordings. The results do not establish junction rectification or its absence. The package preserves original code, source hashes, selected record IDs, aggregate data, all resample gains, graph results, figures and repeat instructions. Original Allen measurements remain external under their source terms. A separate SQL and decimal check and an exact numerical repeat passed.

electrical coupling · gap junction · input resistance · mouse cortex · reciprocity · local recording graphs · reproducibility