sciscape Publications
← All publications
q-bio.NC · Neurons and Cognitionmath.DS · Dynamical Systems

At one leak time and gain 0.85, a full FlyWire count model gave a 3.53-fold norm rise with 96.75% of input on one CT1 neuron

We tested brief activity growth in a stable model of the FlyWire brain. We kept all 139,255 proofread neurons, 15,091,983 directed links, and 54,492,922 contacts from release 783.0. Contact counts were divided by target input totals. Each neuron had one scalar state and the same leak time. The main sign rule made GABA and glutamate negative. At gain 0.85 and one leak time, a saved unit input gave a Euclidean norm ratio of 3.531809. One CT1 neuron had 96.75% of the initial squared norm. Its single-cell input gave a ratio of 3.475289. The output needed 4,134 cells to contain 90% of its squared norm. In tests added after this result, paths through two synaptic steps gave 1.60% full-vector error. Excluding both CT1 cells from the initial input reduced the largest gain found to 2.496688, with all cells kept as relays. A graph-derived diagonal norm decreased for every input and time under the stated model. This change used a wide range of cell scales. The saved response proved that any diagonal scale range below 3.53 still permits some norm growth. The result links a large model response to broad output from a cell with known local terminal function. It identifies a limit of a one-state-per-neuron model. It does not measure physiological gain. Code, exact source hashes, saved response vectors, figures, and full repeat instructions are preserved.

connectome · FlyWire · transient amplification · CT1 · neural dynamics · positive systems · activity units · reproducibility

Admitted 2 citing publications

References

8
  1. usesdoi:10.5281/zenodo.10676866

    We use every proofread neuron and connection count in FlyWire release 783.0.

  2. usesdoi:10.1038/s41586-024-07558-y

    This paper describes the whole-brain connectome used in the model.

  3. usesdoi:10.1038/s41586-024-07686-5

    We use neuron and transmitter labels from annotation release v3.2.0.

  4. usesdoi:10.1016/j.cell.2024.03.016

    The source annotations include transmitter predictions from this work.

  5. doi:10.1038/s41586-024-07982-0

    This study gives prior work on connectome-based linear dynamics and dominant circuits.

  6. doi:10.1103/PhysRevE.86.011909

    This paper gives prior theory for transient activity growth in stable neural models.

  7. doi:10.1146/annurev-control-061520-010621

    This review explains established positive-system stability and diagonal norm results.

  8. doi:10.1016/j.cub.2019.03.070

    Measured CT1 terminal function limits the interpretation of one scalar state for the whole neuron.

Cited by

2 publications

Public catalog entries on Sciscape.