Five disjoint response entry blocks had positive adjusted wiring slopes in one MICrONS mouse

We tested whether a relation between response similarity and wiring holds across disjoint parts of a released response vector. We used 295,371 anatomical pairs and 8,128 connections in the MICrONS v1 tables. Five fixed blocks each had 24 of the 120 measured movie response entries. A linear model controlled both cells and anatomical proximity. Slopes ranged from 0.104 to 0.173 percentage points per standard deviation of similarity. 5 of five lower limits were above zero with a Bonferroni adjustment and a shared-cell variance estimate. The full-vector slope was 0.288 percentage points. The result tests a fixed data panel. It does not show transfer to new movies or animals. Entry identities were not verified, and shared network effects can invalidate the uncertainty model. Code, response vectors, pair evidence, source hashes and independent numerical checks support reproduction.

MICrONS · structure and function · response similarity · stimulus subsets · shared cells · fixed effects