With cell group and session controls, response delay increased by 0.31 ms per 100 micrometres of soma distance in 1,094 mouse V1 connections

We tested whether soma distance carries response delay information after cell group and recording session controls. We used 1,094 detected connections from Allen Synaptic Physiology r2.1 in mouse primary visual cortex. Specimen names gave 356 inferred mice. The common linear delay slope was 0.305 ms per 100 micrometres. Its 95% donor bootstrap range was 0.192 to 0.429 ms. Finer cell groups, separate recording modes, and electrical-pair removal retained a positive slope. Adding distance to a donor-held-out ridge model reduced squared error by 3.35%. Its descriptive fixed-prediction range was -2.37% to +7.83%, so the prediction gain was not precise. A later graph audit found only ten of 511 sessions with cycles in the source-target design. Exact source-cell and target-cell offsets left 15 residual degrees of freedom and distance information in seven sessions. Thus, these data support a group-adjusted distance relation but do not establish a rule independent of individual cell properties. A 48-connection human temporal cortex sample gave no precise slope or prediction gain. Source delays depend on annotated onset values and constrained fits. Soma distance is not axon path length, and the slope is not an axonal conduction speed. The package preserves new code, source hashes, cohort IDs, predictions, graph audits, all results, figures, and repeat instructions. Original Allen records remain external under their source terms.

synaptic latency · cortical circuits · soma distance · paired recordings · mouse V1 · structure and function · donor validation · graph identifiability