At gain 0.85 and one leak time, a full FlyWire v783 model with independent 5 micrometre CT1 states had CT1 input norm ratios below 0.473

We tested how local CT1 states change a whole-brain model response. We kept all 139,255 proofread neuron identities and all 54,492,922 chemical contacts in FlyWire 783.0. We replaced the two CT1 states with 6,276 independent spatial cubes of width 5 micrometres. Each target state divided its incoming counts by its full count. All states had equal leak. At coupling gain 0.85 and one leak time, the largest norm ratio found over all CT1 initial patterns was 0.407559. A positive comparison bounded every CT1 input ratio below 0.472432, with the same count magnitudes and any edge signs. Single whole-cell CT1 inputs in the parent model gave ratios 3.129918 and 3.475289. The input budgets differ in their assumed spatial coherence. When every cube had initial amplitude one, the response norm on the 139,253 other neurons stayed within 0.5% of the whole-cell result. An exact one-step identity confirmed preservation of total outgoing influence. Other grids, sign rules and internal coupling tests also gave ratios below one at the tested time. Cubes are assumed states, not measured electrical terminals. This result is a model benchmark, not a measured change in CT1 gain. The package includes the full count graph, selected contact coordinates, code, input patterns, results, figures, source hashes and repeat instructions.

connectome · FlyWire · CT1 · compartment model · transient response · input coherence · whole brain · reproducible benchmark