In one electric fish map, cell degrees in 10 µm distance groups fixed 56 of 66 reciprocal MG pairs
We tested whether reciprocal links between medium ganglion (MG) cells exceed a reference with cell degrees and soma distance fixed. We used 117 cells and 445 connections between opposite MG types in one electrosensory lobe map. The graph had 66 reciprocal pairs. With each cell's input and output degrees fixed in each 20 µm distance group, 10,000 sampled graphs had a mean of 51.89 pairs. The central 95% reference range was 46 to 58. With 10 µm groups, 412 connections and 56 recorded reciprocal pairs were fixed in every allowed graph. We counted all 34,560 allowed graphs at this resolution. Their mean was 62.83 pairs, and the exact two-sided tail probability was 0.0787. A linear model used equal connection weights, unit cell leaks and normalized inhibitory gain 0.8. Its mean steady response to input at the same cell was 1.0617 for the recorded graph. Reference means were 1.0491 for 20 µm groups and 1.0590 for 10 µm groups. The recorded response was inside the exact 10 µm reference range. Fine constraints retain most reciprocal wiring by construction. Thus a smaller excess at this resolution does not show that soma geometry alone explains the wiring. We provide the graph, source checks, exact graph spaces, sampling code, model results and repeat instructions. The study does not measure learning rate or variation across fish.
connectome · electric fish · electrosensory lobe · medium ganglion cells · reciprocal connectivity · spatial controls · inhibitory feedback · graph sampling
References
5- usesdoi:10.5281/zenodo.19892261
This versioned dataset supplies outgoing synapse rows, cell types and reconstructions.
- usesdoi:10.1038/s41586-026-10690-6
This study supplies the circuit context, cell type meanings and data interpretation.
- usesdoi:10.1137/16M1087175
This work explains graph references with fixed degrees and the need to define their sample spaces.
- usesdoi:10.1093/biomet/57.1.97
The sampler uses a Metropolis correction for unequal proposal probabilities.
- doi:10.1016/S0925-2312(01)00481-7
Prior modelling linked mutual inhibition to adaptation rate. We test a different fixed response model and do not measure adaptation.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.