Fruit Fly Lab

Other organisms

The organism is a swappable component. The worm's whole nervous system runs here too. Beyond the worm, this page shows honestly what the same experiment would need in a fish, a mouse and a human, and what is missing.

The worm, running in your browser

C. elegans has 302 neurons; this wiring diagram covers 279 of them (Varshney et al. 2011, WormAtlas). Most worm neurons do not spike, so the brain-model component changes too: a graded-voltage model (Kunert, Shlizerman & Kutz 2014, parameters from Kunert-Graf et al. 2017). The test is the classic touch reflex: touch the tail and a real worm moves forward, touch the head and it backs up (Chalfie et al. 1985). The readout is the balance between the forward command neurons (AVB) and the reversal command neurons (AVA).

The standard sign assumption treats every non-GABA synapse as excitatory. The touch neurons release glutamate, and in the worm some glutamate synapses inhibit through glutamate-gated chloride channels. With the standard assumption the model gets the tail-touch reflex wrong. Making all of the touch neurons' outputs inhibitory (the second sign option, a hypothesis rather than a dataset) does not rescue it either. That is the point: a complete wiring diagram is not enough; you also need to know what each connection does. A whole-brain recording atlas now tests this directly and finds that signals often do not flow where the anatomy predicts (Randi et al., Nature 2023; data: DANDI 001075).

The ladder, worm to human

What this lab does for a fly (full wiring diagram, a model that runs on it, a behavior to check against) exists today only for the smallest brains. Numbers below are from the cited papers.

OrganismNeuronsMappedRunnable whole-brain modelWhat this experiment would still need
Worm (C. elegans)302Whole animal (Cook 2019)Yes, aboveWhich connections excite and which inhibit
Fruit fly larva3,016Whole brain, 548,000 synapses (Winding 2023)PartialMatched behavior recordings
Adult fruit fly139,255Whole brain, ~50 million synapses (Dorkenwald 2024); male brain + nerve cord 166,700 (Berg 2026)Yes: the lab (Shiu 2024)The nerve cord in the same animal; recordings during tasting
Larval zebrafish~100,000Partial; whole-brain EM being proofreadNoA finished wiring diagram. Activity recordings of >70,000 neurons already exist (ZAPBench)
Mouse~71 million1 mm³ of visual cortex: >200,000 cells, 0.5 billion synapses (MICrONS 2025), well under 1% of the brainNoThe other 99%+ of the wiring
Human~86 billion1 mm³ fragment: ~57,000 cells, ~150 million synapses (H01, 2024), about a millionth of the brainNoAlmost everything

What carries over to mammals, and what does not

So the honest extrapolation is a method, not a result: map the wiring, run it, intervene, and check against a real animal. Each rung up the ladder is gated on the same missing pieces.