Winner-Take-All Network

Competitive inhibition, decision making, and neural selection

Inhibition Strength 50
Input Noise 20
Decision Threshold 60
Input Bias 10
Winner
Decision Time (ms)
0
Decisions Made
Coherence

About

The winner-take-all (WTA) network implements neural competition: a pool of excitatory neurons each receive distinct inputs and all inhibit each other through shared interneurons. The neuron receiving the strongest input quickly suppresses all others through this lateral inhibition, achieving a winner state that is robust and stable. WTA dynamics underlie cortical decision-making (Wang, 2002) — the decision is complete when one "attractor" state crosses threshold, a process that takes longer with weaker input differences (Hick's law) and more noise (speed-accuracy tradeoff). The same mechanism implements categorical perception, attention, and motor selection in the basal ganglia.