With additive exponential tails, second voltage products stayed lower in 56 reciprocal mouse V1 pairs at 30 Hz

We tested whether remaining first-response voltage explains a smaller second response product in reciprocal PV and pyramidal cell connections. We used public 30 Hz slice recordings from mouse V1. An additive model subtracted a fitted exponential first-response tail at each source second peak. The primary sample had 56 pairs from 25 mice. The equal-mouse mean bounded product change was -0.300, with a 95% mouse resample range of -0.361 to -0.241. The same pairs gave -0.321 with local baselines and -0.037 with the original baseline. The paired model-minus-local shift was uncertain. All specified checks retained a decrease, but a linear continuation stress test gave a smaller estimate of -0.092. A withheld six-ms segment showed positive inhibitory forecast bias of 0.054 first-amplitude units. The model therefore did not fully capture the observed first decay. Odd and even trial scores had modest pair agreement. Independent solvers and exact-model synthetic checks passed. These results describe the tested tail assumptions. They do not identify an isolated second response, conductance change or feedback gain. The package preserves code, full and split waveforms, source hashes, all fits, scores, resample draws, figures and repeat instructions.

reciprocal connections · short-term plasticity · voltage overlap · PV interneurons · mouse visual cortex · model validation