PSD weighting reduced the targeted deletion margin in a pooled sample of 155 ferret V1 spines

We tested how weight spread affects a fixed orientation projection in a public table of 155 ferret V1 spines. We defined the margin as the minimum number of row deletions that makes the projection nonpositive. Sorting signed contributions gives an exact solution. Equal weights required 36 deletions. PSD area weights required 19, although the initial projections were similar, at 0.222 and 0.215. Square-root PSD weights required 25. All observed weighted counts were within central 95% ranges from 20,000 random weight assignments. None of 20,000 uniform random deletions at each targeted threshold erased alignment. These results show that similar pooled alignment can hide different targeted deletion margins. They do not require a special size and tuning relation in this reference. The table lacks cell and animal IDs. The result is a finite-sample assay, not a lesion prediction or a claim of transfer across animals. We provide source hashes, code, exact certificates, random reference counts, and a figure.

dendritic spines · structure and function · orientation tuning · targeted deletion · weight spread · ferret V1