Role of synaptic noise in synchronization-avoiding critical dynamics

Determine how introducing synaptic noise affects the dynamics of the all-to-all network of KTH map neurons with diffusive gap-junction coupling and the local anti-Hebbian homeostatic plasticity rule W_{ij}[t+1] = W_{ij}[t] + (1/τ)(A − W_{ij}[t]) − U_W W_{ij} S_i[t] S_j[t], specifically its self-organization toward the edge of the synchronization transition, the maintenance of near-critical operation (W* ≈ W_c), and the optimal input reverberation characterized by critical slowing down.

Background

The paper introduces a network of KTH map neurons coupled via gap junctions that undergoes a continuous synchronization phase transition at a critical coupling W_c. Near W_c, inputs reverberate optimally due to critical slowing down.

A local anti-Hebbian homeostatic plasticity mechanism is proposed that depresses synapses upon spike coincidence and otherwise recovers weights toward a baseline A. This Synchronization-Avoiding Self-Organization (SASO) robustly hovers near the synchronization critical point without requiring an absorbing state.

While the model explores robustness across system size and hyperparameters, the influence of synaptic noise on maintaining quasi-critical dynamics and on reverberatory processing is left unresolved and explicitly posed as an open question.