Neural Field Bump Attractor

Network Parameters

N neurons 256 Excitatory width σ_E 0.15 Inhibitory width σ_I 0.35 Excitatory strength J_E 2.0 Inhibitory strength J_I 1.8 Time constant τ 10 Noise σ 0.02

Input

Input position 0.5 Input strength 0.0

Observables

Bump position—

Bump amplitude—

Bump width—

Total activity—

τ du/dt = −u + W*f(u) + I

Mexican-hat connectivity W(x) = J_E·G(σ_E) − J_I·G(σ_I) supports a localized "bump" state. This models spatial working memory: bump position encodes remembered stimulus location and persists without input. Input shifts the bump; noise causes diffusion.

NEURAL FIELD BUMP ATTRACTOR

Network Parameters

Input

Observables