Connectome OS
Rust LIF runtime and debugging interface for stimulating, cutting, and measuring FlyWire-derived graph activity. Alpha research preview.
Neurons are simulated — leaky integrate-and-fire or better — with documented, tuned parameters.
Dispute this rating →The evidence
No evidence, no level.The rated repository contains only documentation and points to the implementation in ruvnet/RuVector.
https://github.com/ruvnet/Connectome-OS/blob/d54a5270227ca91abe69a92ac9e9d813769f3954/README.md#L15-L17The LIF kernel defines documented per-neuron biophysical parameters including membrane time constant, threshold, refractory period and reversal potentials.
https://github.com/ruvnet/RuVector/blob/8667f1f4e57055d063f080dfa60c80dd50874d20/examples/connectome-fly/src/lif/types.rs#L7-L46Spike emission reads the stored synaptic weight to schedule delivery; no weight is written during simulation.
https://github.com/ruvnet/RuVector/blob/8667f1f4e57055d063f080dfa60c80dd50874d20/examples/connectome-fly/src/lif/engine.rs#L380-L412The only synaptic weight change is an experimenter-specified lesion that zeroes nominated edges into a perturbed copy before re-running.
https://github.com/ruvnet/RuVector/blob/8667f1f4e57055d063f080dfa60c80dd50874d20/examples/connectome-fly/src/lesion.rs#L127-L145
The ruvnet/Connectome-OS repository itself holds only a README and two screenshots; the README points at the actual implementation, which lives in ruvnet/RuVector on branch research/connectome-ruvector under examples/connectome-fly. Read there, it is a genuine conductance-based LIF runtime with documented per-neuron biophysics (tau_m, v_rest, v_reset, v_thresh, r_m, tau_refrac, excitatory/inhibitory reversal potentials and separate synaptic decay constants), event-driven over FlyWire Princeton CSV graph data, so voltages evolve under a documented neuron model: L1. Not L2: the crate exposes no plasticity module (lib.rs declares analysis, audit, connectome, embodiment, lesion, lif, observer, stimulus) and the spike-emission path only reads s.weight. The single place a synaptic weight changes is Connectome::with_synapse_weights_zeroed, which builds a perturbed copy for an experimenter-nominated lesion before re-running the stimulus - a cut made by the user, not a biological rule running during the episode. RuVector's monorepo does contain STDP code, but in unrelated crates that connectome-fly does not depend on.
Record
Nearby
All projects →Research code for the Shiu et al. connectome-based leaky integrate-and-fire model, including activation/silencing experiments, notebooks, and FlyWire data configuration.
FlyWire whole-brain LIF implementation based on Shiu et al., with multiple simulation backends and benchmarking tools. This repository supplies the neural model; it is not a complete embodied demo package.
Ratings on thelearningfly.com are proposals, not verdicts. Every one of them is arguable in public.