ARBOR
- A browser simulation where neurites grow by space colonization, form synapses on contact, and carry signals across the network they build.
- Status
- Shipped
- Date
- Jul 2026
- Role
- Developer
About
Each run starts from scattered seed points and a field of attractors. A growth tip averages the direction toward the attractors inside its radius, steps that way, and consumes any it reaches. When it runs out of attractors it stops or splits in two, so the branches keep pushing into empty space. If two tips from different neurons pass within a set distance, both stop and a synapse forms where they met; a spatial grid keeps that check fast with up to 300 neurons growing at once. The synapses turn the neurons into a graph, and a signal walks it outward from one neuron with a delay proportional to the length of each segment. You set the neuron count and growth speed, fire a pulse on demand, and toggle the attractor field to see what steers the growth. Reset and you get a different network from the same rules. ARBOR approximates growth and signalling with procedural rules and leaves out membrane potentials and ion channels. It runs on p5.js with no build step and deploys to GitHub Pages.