With uniform source density, two reach scales gave no clear prediction gain for 117 mouse entorhinal recording pairs

Can measured axon and dendrite spans explain the spatial range of shared functional input? We tested an interval model with 117 mouse entorhinal recording pairs from 72 dates and morphology records for 16 stellate cells and 24 parvalbumin interneurons. The model uses uniform source density and fits a baseline and gain for each assay group. Every test excludes its recording date and repeats scale selection. Separate reach scales reduced peak current correlation prediction error by 5.0% compared with one scale. The central 95% range from 500 date resamples, with morphology resampling and all fits repeated, was -16.1% to +25.7%. The fitted inhibitory-to-excitatory scale ratio was 0.440, with a resample range of 0.216 to 1.919. A linear distance model had similar error. A source field within 100 micrometres of the pair midpoint gave a two-scale error decrease of 1.8%. The data did not resolve a prediction gain from separate scales in this model. This does not establish equal biological scales. Dates are not known animal identifiers, and the interval model omits synaptic density and strength. We provide the exact geometry method, source tables, code, predictions, resamples and figures.

entorhinal cortex · structure and function · shared input · axon reach · dendritic geometry · parvalbumin · stellate cell · prediction · reproducible research