In passive cable models of 668 Allen neurons, apparent membrane capacitance was 1.6 to 3.6 µF/cm² and fell with cell size; size explained most of the 0.58 to 0.63 human-to-mouse ratio

We made a passive cable model of each Allen Cell Types Database neuron that has a dendrite reconstruction (668 cells: 158 human, 510 mouse). In each model, the specific membrane resistance gives the measured input resistance. The ratio of the measured membrane time constant to the model time then gives an apparent specific membrane capacitance (Cm). The apparent Cm was much larger than 1 µF/cm²: group geometric means were 1.6 to 3.6 µF/cm², and only 6 to 26 of 668 cells were at or below 1 in seven model settings (axial resistivity 70-300 Ohm cm, axon in or out, fit window, soma shunt, radius floor). Allen all-active model fits of 107 mouse cells also gave a high area-weighted Cm (3.4 µF/cm²), with weak cell-to-cell agreement. In an exploratory test, a somatic leak did not bring Cm to 1 in 608 of 668 cells. Apparent Cm fell with reconstructed area inside each species and cell group (log-log slopes -0.17 to -0.95), which a uniform passive membrane does not predict. Without size control, the human-to-mouse Cm ratio was 0.58 (95% donor-bootstrap interval 0.51-0.67) for aspiny cells and 0.63 (0.53-0.75) for spiny cells, near the ratio of about 0.5 reported by Eyal et al. (2016). Human cells are larger. With area matching, the aspiny ratio became 0.69 to 0.81 (intervals 0.60 to 0.96) and the spiny ratio 0.89 to 1.01 (intervals include 1). Thus, in these data, most of the low apparent human Cm follows from a size relation that also occurs inside each species. Model-based Cm estimates from somatic recordings and light-microscope trees need a size-bias check before they can show a species difference. Limitations: unknown axial resistivity, unreconstructed spines and thin processes, radius errors, passive model with two summary values per cell, few human aspiny cells (38), and area-matching and shunt analyses added after the first results. Allen source data are not included; scripts and hashes allow retrieval.

specific membrane capacitance · cable theory · passive neuron model · dendrite morphology · human neurons · mouse neurons · Allen Cell Types Database · input resistance · membrane time constant · species comparison · size bias