With matched windows, response timing cut error by 0.8 to 1.8% in fixed ridge tests of 52 mouse V1 inhibitory connections

We tested whether visual response timing adds synaptic strength information in mice excluded from model fitting. We used open paired recordings and visual calcium responses from mouse primary visual cortex. Extraction reproduced the source cohorts and strength correlations: 52 inhibitory connections from 18 mice, r = 0.434, and 40 excitatory connections from 19 mice, r = 0.389. A fixed ridge model predicted log voltage magnitude from cell distance and time-averaged response similarity. Adding full response similarity cut equal-mouse error by 11.46% for inhibition. Its 95% mouse bootstrap range was -19.51% to +39.23%. The source comparison used different response windows for the two features. With the same full-trial frames, the gain was 0.83% (-34.21% to +32.36%). With the same stimulus-on frames, it was 1.80% (-22.85% to +26.35%). Excitatory matched-window gains were negative. Paired ranges also leave the decrease in gain uncertain across mice. These results do not establish a timing benefit or its absence. They show that frame selection affects this strength prediction test. The package provides a benchmark, derived response arrays, pair tables, complete code, all results, figures, source hashes, and reproduction steps. The small, selected sample and voltage response measure limit biological claims. The window bootstrap checks were added after the planned point estimates were known.

synaptic strength · visual cortex · PV interneurons · structure and function · response timing · cross-validation · window selection · paired recordings