Sixteen schematic brain regions (prefrontal, motor, parietal, occipital, temporal, hippocampus, amygdala — left/right — plus thalamus and brainstem as central relay hubs) sit in a fixed 2D layout, joined by a weighted structural-connectivity graph loosely modelled on real cortico-thalamic and callosal pathways.
Each node is a leaky integrate-and-fire unit. Its membrane potential V decays exponentially toward zero between inputs:
dV/dt = −λ·V
fires when V ≥ θ (then V → 0, node enters refractory period τ)
When a node fires it sends a pulse along every active edge (weight ≥ density cutoff). The pulse is not instantaneous: it arrives at the neighbour after a delay proportional to the on-screen distance between the two regions divided by the propagation-speed slider — exactly like conduction delay along a real axon bundle. On arrival, the neighbour's potential increases by edge weight × pulse gain, which can push it over its own threshold and continue the cascade.
- Threshold θ — how much accumulated input a region needs before it fires; raise it and cascades die out quickly, lower it and one stimulus can sweep the whole network.
- Propagation speed — sets conduction delay; slow speeds let you watch the wavefront travel edge by edge.
- Connection density — prunes the graph to only the strongest structural edges (low) or keeps the full connectome (high), changing which regions a cascade can reach.
- Leak / decay — how fast unfired input dissipates; high leak means inputs must arrive close together in time to summate.
- Click any region on the canvas to stimulate it directly.