Can We Digitize a Neuron? (www.these.us)

Researchers argue that C. elegans — a tiny, transparent worm whose entire nervous system (302 neurons) was first reconstructed in 1986 — is uniquely positioned to answer the question “Can we digitize a neuron?”. The original connectome required ultrathin (~50 nm) sectioning, electron-microscopy imaging and decade-long manual tracing to reveal ~5,000 chemical synapses, ~600 gap junctions and ~2,000 neuromuscular junctions. Today we also have its genome and single-cell RNA-seq, giving molecular identities for every neuron. Together, morphology, connectivity, molecular profile, electrophysiology and biophysical models are the ingredients needed to create a faithful digital twin of a neuron — but each ingredient matters: synaptic multiplicity and graded strengths, neuromodulators that reconfigure circuits, and individual variability complicate a one-to-one mapping. Practically, optogenetics (opsins like Channelrhodopsin-2 and Halorhodopsin) supplies a reproducible experimental handle to activate or silence identified cells with millisecond precision, letting scientists close the loop between stimulation, behavior and recorded activity. For AI/ML, a digitized neuron in C. elegans would provide the rare ground-truth, causal dataset linking structure, molecular state and function — ideal for validating biophysical and neuromorphic models, discovering biologically informed learning rules, and training ML systems that incorporate realistic priors about computation and plasticity. The worm thus offers a tractable, high-resolution testbed to move from wiring diagrams to mechanistic, testable simulations of single neurons and small circuits.