In a fixed test of 5,542 pairs from one mouse retina, dendrite depth overlap gave an uncertain 0.21% error decrease for residual calcium correlation

We tested whether overlap of dendrite depth profiles helps predict shared variation in calcium responses. The e2198 panel had 315 matched ganglion cells in 47 types. We rebuilt all 634 exported traces from nine original recordings with zero error. Only 5,542 of the 49,455 possible pairs had final traces from the same recording. Pair scores from two disjoint unfiltered repeat contrasts had Pearson correlation 0.714. A nested ridge test excluded both cells of each test pair from fitting and penalty selection. Controls included spatial distance, planar arbor overlap, visual response similarity, response quality, and type. Adding depth overlap reduced squared test error from 0.02062917 to 0.02058545, or 0.212 percent. A fixed-fit central 95 percent cell-weight range was -0.331 to +0.768 percent. A recording-resample range was -0.123 to +0.691 percent. Both ranges included zero in all 14 model cases. The result gives no clear added gain for this scalar depth rule under the tested controls. It does not establish no anatomical effect. One retina, four retained repeats, calcium measurement, possible optical mixing, and ranges without model refitting limit inference. The package preserves a verified recording map, compact numerical data, code, provenance, predictions, figures, and source rebuild instructions.

retina · structure and function · calcium imaging · shared response variation · dendrite depth · reproducible benchmark