NEXT BRAIN Computational Neuroscience & NeuroAI
Models/FlyWire · MNIST Readout
Connectome-constrained taskBrian2-style model / Rust and WASM runtime

FlyWire · MNIST Readout

A full FlyWire-based network with fixed neural dynamics and an external digit classifier.

FlyWireMNISTConnectomeWhole Brain
Model backend supportFull network · 139,255 neurons

Model support is separate from currently registered workers · activity across all scales

BrowserWASM runnableActivity unavailable
Single CPUUnavailableActivity unavailable
Single GPUUnavailableActivity unavailable
DistributedUnavailableActivity unavailable
Local activityBrowser reports are client-reported
—Browser reports
—Compute runs
—Forks
—Shared links
—Queued now
Research status · research result70,000-image campaign completed; fixed-network activity reached 87.14% external-readout test accuracy.

The full simulation runs through the dedicated FlyLab browser interface. The classifier is a separately trained external readout; CPU, GPU and distributed queue adapters are not available.

Open standalone FlyLab ↗
Execution

Run this model

Choose a model size and computing resource when you're ready.

Scientific overview

What this model represents

The research experiment retained all 139,255 FlyWire v783 neurons and 15,091,983 directed weighted connections representing 54,492,922 synaptic contacts. Artificial image spikes entered through 302 antennal-lobe projection neurons; an external readout decoded fixed downstream activity. The packaged FlyLab browser worker runs the full network locally with the f64 WebAssembly engine.

Neuron model
Simplified spiking neurons
Biological system
Drosophila whole brain
Diagram nodes
1
Largest listed scale
139,255