The somatosensory cortical map — Penfield's homunculus — is not printed by genes. It emerges from competition. Each cortical neuron wires itself to the input it receives most, and neurons adjust continuously based on which inputs are active.
This simulation runs 60 neurons competing over 6 input channels (representing body-surface regions). Each neuron updates its preference toward whatever it receives — and updates its neighbors too. The map that forms reflects current input. Remove an input, and its territory gets absorbed by neighbors. Restore it, and the competition resumes.
Toggle channels on and off. Watch the colored territory redistribute. The mechanism is the same one that lets TVSS users eventually perceive camera space rather than tongue stimulation: not special plasticity, just competition running on new input.
What it can't show: real cortical reorganization operates at two timescales. Fast changes (minutes to hours) come from unmasking of latent synaptic connections — the map can expand into a neighboring region immediately after input removal because silent synapses were already there. Slow changes (days to weeks) require axonal sprouting. This simulation compresses both into the same learning rule.
Also: the ordinal structure of this 1D strip (thumb–index–middle–ring–pinky–palm) is approximately right for the hand region of somatosensory cortex, but the actual human map is 2D, non-linear, and varies across individuals. Adjacent channels in this simulation tend to cluster because of a mild correlation in their input distributions — a shortcut for representational proximity that, in the brain, comes from the geometry of the body surface itself.