Simple circuits can create complex behaviors

A virtual creature called a “spark” demonstrates how simple neural wiring can produce surprisingly varied behavior. Its basic parts are sensory neurons that detect light and motor neurons that move flippers. With one eye and one flipper, it can speed up in brighter conditions but cannot steer. Adding a second eye and flipper allows steering, but directly connecting each eye to the flipper on the same side makes the spark turn away from light. Crossing the connections reverses its direction and draws it toward the light, though it then rushes past rather than stopping. Changing the connections from excitatory, which increases movement, to inhibitory, which reduces it, lets the spark turn toward the light and slow to a stop.

These small circuit changes yield distinct apparent temperaments: attraction, avoidance, aggression, or a more measured approach. The demonstration argues that behavior that might look intelligent can emerge from a few neurons, without any hidden complexity or magic. It also shows how differences in connection strength give otherwise similar sparks different preferences, much as brain wiring may shape individual likes and dislikes. The model is deliberately simplified: real brains contain vastly more neurons, and unlike these sparks’ fixed circuits, they can change through learning. The examples build on thought experiments about simple artificial organisms and set up the idea that learning involves rewiring neural connections.