Visual motion decision · Connectome Lab

Experiment0.000 s

Visual motion decision. A configuration plus a seed fully determines the run.

ControllerCOUPLING OFF

Action selection produced by a simulated network whose connectivity is measured Fish1 anatomy.

The animalONE MODEL

The body is MODELED reference anatomy. The brain inside it is measured.

Traced morphologyMEASURED

Loading reconstructed morphology…

Motion

Coherence60%

Speed0.60 rad/sDuration12 sNoise0.00

Loop

The fish's own rotation subtracts from the retinal motion, so turning toward the pattern reduces the evidence driving the turn.

Seed42

In silico ablationMODEL ONLY

Removes the one MODELED pathway. Tests how much the behaviour rests on it.

Removes the 300 MEASURED recurrent pairs that produce the integration.

Silences one side entirely. Decisions in one direction become impossible.

Silences the measured descending output the decision is read from.

Ablations change the MODEL, not an animal. Re-run with the same seed to compare against the control run.

World / fish—

STIMULUS SIMULATED

Brain / HMI

POSITIONS MEASUREDACTIVITY SIMULATED

Evidence

L 0.00 · R 0.00

simulated sensor

Decision

—

integrating

Action

NONE

—

Source

PROCEDURAL

no connectome

Neural step

0.00 ms

200 Hz fixed