Finite rise left the mean timing-penalty change unresolved in selected Allen cortical input models

We tested finite-rise and instantaneous-rise linear cancellation models for excitatory and inhibitory inputs to one target in Allen slice recordings. We matched passed current-clamp fits by target, experiment and holding voltage. The mouse visual-cortex sample had 73 E/I pairs and 51 inferred donors. The human temporal-cortex sample had 10 pairs and five inferred donors. The mean finite-minus-instant timing-penalty difference was -0.0053 (-0.0252 to 0.0128) in mouse and 0.0091 (-0.0201 to 0.0396) in human, in units of excitatory response energy. Intervals are nominal 97.5% donor bootstrap intervals. Neither sample resolved a mean change. This does not establish model equivalence or measured biological cancellation. The models assume free inhibitory gain, pair-specific timing control, linear addition and rise power two. Fit uncertainty, selected slice circuits and small human samples limit inference. New code, aggregate results, a figure, source hashes and external-data regeneration instructions support inspection. External records are not bundled.

Allen Institute · synaptic rise time · excitation and inhibition · cancellation · convergent inputs · conditional model · null result