Continuous attractor dynamics in a 2D neural field (ring attractor model)
Click canvas to move stimulus — bump tracks input position
Neural field equations (Wilson & Cowan 1972; Amari 1977) describe spatially continuous neural populations. With Mexican-hat connectivity (local excitation, surround inhibition), stable "bump" attractors emerge that encode continuous variables like head direction or working memory content. The bump can be driven by external input or drift spontaneously along a continuous attractor manifold.