Mean voltage bounds resolved response order in 22 of 48 mouse V1 pairs at 30 Hz under nonnegative monotone additive tails

We tested which comparisons of second responses require a fitted first-response decay. We used 48 reciprocal PV and pyramidal cell pairs from 22 mice. Inputs were about 33 ms apart. We rebuilt 1,852 trials from the raw source. The model assumes additive responses and nonnegative first tails that cannot increase. At measured means, bounds placed normalized second inhibition below excitation in 21 pairs and above it in one pair. Sixteen pairs remained unresolved. Ten pairs had negative reference means and received no finite decision. With a five-trial block resample, only three pairs retained a 97.5-percentile upper bound below one. This was an uncorrected per-pair threshold. The observed equal-mouse relative ratio was 0.516 with source local baselines and 0.826 with the first baseline. Their paired ratio was 1.599, with a 95% mouse resample range of 1.330 to 1.859. Earlier SCI work already established the distinction between local and total voltage products. We add scalar tail bounds, pre-spike reference measurements, and a trial uncertainty audit. We moved the main reference from 1 ms to 5 ms before the second spike after trace inspection. This exploratory change is preserved. The results depend on the component model and fixed source selection and peaks. They do not identify conductance, release probability, or intact feedback gain. Code, source hashes, all used data, bounds, resample draws, figures and repeat instructions are preserved.

functional connectivity · postsynaptic potentials · response overlap · partial identification · short-term plasticity · inhibitory circuits