#15 · Systems biology

Inhibitory integration, burst-driven weight attractors, and Φ in spiking networks

The dominant thread is how network architecture and collective activity constrain neuronal input-output transformations and synaptic plasticity in simulated spiking networks. A derivation of probabilistic rules for feedforward inhibitory circuits describes how excitatory and inhibitory inputs are integrated to shape stimulus-evoked responses. A network model shows that collective bursting pulls synaptic weights into an analytically derived attractor, enabling selective consolidation. A separate application of IIT 3.0 to integrate-and-fire networks reports non-zero Φ and sensitivity to noise, a less directly circuit-level result. Together these studies treat input integration, weight dynamics, and integrated information within simulated spiking models. What is new is the analytical treatment of excitatory-inhibitory balance and attractor structure in weight dynamics, alongside an information-theoretic measure applied to the same class of network.