Local radius scores exceed finite passive loads at sampled cuts in selected Allen dendrites

We tested whether local radius scores establish finite downstream load in passive DC dendrite models. We used public Allen Cell Types reconstructions and sampled three arms 5 micrometers from each basal branch. The primary sample had 2,450 branches in 210 mouse spiny cells from 158 donors and 1,622 branches in 91 human spiny cells from 28 donors. With axial resistivity 150 ohm cm and membrane resistance 20,000 ohm cm squared, equal-donor geometric mean local scores were 1.588 and 1.592. Finite downstream tree scores were 0.257 and 0.376. The donor-averaged fractions of branches with local score above one but finite score below one were 82.9% and 88.3%, with pointwise 95% donor bootstrap intervals of 80.8-84.9% and 84.8-91.6%. A majority changed direction at membrane resistance 5,000 and 80,000 as well. These are loads at daughter cuts normalized by parent characteristic conductance, not measured common-junction conductances. Assumed passive parameters, sealed recorded tips, radius error, and selected tissue limit inference. The package has code, derived branch evidence, exact source hashes, independent cable-network checks, and retrieval instructions.

Allen Institute · dendritic computation · Rall rule · finite cable · passive conductance · boundary conditions