Basal tree asymmetry gave no clear adaptation prediction gain in selected Allen mouse and human spiny cells

We tested whether basal dendrite branching asymmetry improves prediction of stored spike-frequency adaptation after tree size, passive electrical properties, and broad cell labels are known. We linked reconstructions and electrical summaries from the same Allen cells. Donor tests used nested ridge models and three fold partitions. In 206 mouse spiny cells from 157 donors, added asymmetry changed mean squared error from 0.278745 to 0.279197 on an asinh-transformed adaptation scale. The error gain was -0.000452, with a nominal 97.5% donor interval from -0.005598 to 0.004786. In 69 human cells from 23 donors, the gain was -0.003891, with an interval from -0.062598 to 0.064398. Neither test gave a clear gain. Planned secondary checks and transfer between species also gave no clear gain. Intervals resample fixed prediction errors and omit model-refitting uncertainty. Missing adaptation values, cut arbors, small samples, clinical tissue, and linear models limit the result. This is not a test of all dendritic computation or a causal mechanism. The package provides a topology benchmark, source hashes, code, derived evidence, a figure, and independent numerical checks.

Allen Institute · dendritic topology · spike-frequency adaptation · donor validation · structure and function · null result