Self-consistent neuronal population under spike inputs and unbalanced conditions

A single neuron gain function can predict the population activity of homogeneous neurons under strong limitations, such as the stationary state and balanced conditions of the total input. In this work, we propose a modification to the self-consistency model when balanced conditions are not fully satisfied. We present a scaling factor to modify the excitatory weights in a Brunel network. It allows using the self-consistency model in more realistic cases. The approach is used and analyzed for different network features.