In two mouse V1 projects, inhibitory onset delays were 0.23–0.37 ms shorter within 230 two-way cell pairs

We compared synaptic onset delays within cell pairs that had one recorded excitatory direction and one inhibitory direction. We used Allen Synaptic Physiology r2.1. The mouse V1 coarse-matrix sample had 86 pairs from 53 inferred donors. Inhibitory minus excitatory delay had mean -0.228 ms and a 95% donor resample range of -0.327 to -0.124 ms. A separate patch-seq project had 144 pairs from 103 donors. Its mean was -0.369 ms, with a range of -0.429 to -0.309 ms. No donor was shared between projects. Current-clamp, voltage-clamp, annotated-onset, and known-cell-class checks retained negative contrasts. Many fits reached an onset constraint, so annotation and fit error remain limits. A later test used the first project mean to predict the direction difference in the second. Squared contrast error fell by 34.6%, with a two-sample donor resample range of 21.6–46.5%. This test was added after the main results were known. Equalizing each pair at its mean delay gives median delay-only phase errors of 2.31 and 2.46 degrees at 40 Hz. It preserves the total delay and stability in a stated two-cell scalar model. These slice results do not establish conduction speed, intact-circuit firing time, or transfer to another laboratory. The package preserves original code, source hashes, the selected numerical audit, all resamples, figures, and repeat instructions. A full source repeat and independent SQL checks passed.

synaptic delay · reciprocal connections · excitation and inhibition · mouse visual cortex · paired recordings · structure and function · reproducible analysis