Descending-neuron atlas

More than reference material: besides the wiring-reach map, the repo ships its own leaky integrate-and-fire simulator over the MaleCNS v1.0 brain + nerve cord, explicitly in the style of Shiu et al. (2024), with a stated parameter set (v_rest -52 mV, v_th -45 mV, tau_mem 20 ms, tau_syn 5 ms, 2.2 ms refractory, 1.8 ms axonal delay), Poisson drive of the selected descending neurons at 75 or 150 spikes/s, and membrane voltages integrated at dt = 0.25 ms. Membrane voltages change under a documented neuron model, so L1. Not L2: no synapse is ever updated - every synapse carries the same fixed W_SYN = 0.08 mV signed by the presynaptic transmitter, hand-calibrated once by the author, and a grep for plasticity, learning, STDP or dopamine-driven weight change across the Python and JavaScript finds nothing. Worth carrying over that the project is unusually candid about this: its own Limitations section says the model has no gap junctions, no neuromodulators, no sensory feedback and starts from rest, and that actions show muscle recruitment rather than movement. Caveat for a reader: the shipped page replays precomputed results from sim.js; the integration itself runs offline in simulate.py at build time.