Cell-body distance improved delay predictions after PSP rise time in 1,136 Allen mouse inhibitory connections

We tested whether cell-body distance adds prediction value for synaptic onset delay after the rise time of the postsynaptic potential (PSP) is known. We used linked geometry and physiology records in the Allen Synaptic Physiology r2.1 small database. Ten repeated five-fold partitions held out whole inferred donors. Models used source and target cell labels and measured PSP rise time. In 1,136 mouse inhibitory connections from 469 inferred animals, distance reduced donor-weighted RMSE from 0.4879 to 0.4653 ms. The mean squared error gain was 0.02160 ms squared, with a conditional 98.75% donor bootstrap interval of 0.00370 to 0.04387. The gain remained positive with other penalties, raw rise time, and joint cell labels. Mouse excitatory and both human groups had uncertain gains. Measured rise time did not remove the distance signal in the selected mouse inhibitory sample. This does not identify conduction speed or a causal mechanism. Slice sampling, coupled response fits, missing measurements, inferred donor labels, and uncertainty that holds fitted models fixed limit the result. Code, selected records, all held-out predictions, a figure, and an independent evidence check are supplied.

Allen Institute · synaptic delay · rise time · cell-body distance · donor validation · inhibitory connections