Reverse current gave no clear added strength prediction in 69 reciprocal mouse V1 pairs with test mice excluded

We tested whether a known excitatory current helps predict the inhibitory current in a reciprocal cortical cell pair. We used Allen Synaptic Physiology r2.1. The primary sample had 69 mouse V1 pairs from 53 mice. Each test mouse stayed outside model fitting. A fixed ridge model controlled cell labels, geometry, age and recording project. Adding the reverse current increased squared log error by 4.64%. The central 95% animal resample range for error reduction was -13.58% to +12.25%, with model refits. An overlapping 141-pair selection check gave +0.021%, with range -4.80% to +7.53%. A later nested model-selection check also gave no clear gain. Only 34 primary pairs had both voltage and resistance data. They did not establish a membrane explanation for earlier voltage-matching results. Only eight pairs passed the chosen current-fit residual rule in both directions. These findings limit the tested weight-completion rule. They do not show independent synaptic strengths or reject matching in another tissue sample. The package preserves source code, cohort IDs, source hashes, all predictions, resample draws, figures and repeat steps. The original database remains external under its source terms.

reciprocal connections · synaptic current · structure and function · mouse visual cortex · prediction · input resistance