Training recurrent spiking neural networks in strong coupling regime

Abstract

<jats:title>Abstract</jats:title><jats:p>Recurrent neural networks can be trained to perform complex tasks. However, due to the large number of unconstrained synaptic connections, the recurrent connectivity that emerges from network training may not be biologically plausible and thus it is questionable as to whether they are applicable to actual neural processing underlying these tasks. To narrow this gap, we developed a training scheme that, in addition to achieving learning goals, respects the structural and dynamic properties of a standard cortical circuit model, i.e., strongly coupled excitatory-inhibitory spiking neural networks. By preserving the strong mean excitatory and inhibitory coupling of initial networks, we found that trained networks obeyed Dale’s law without additional constraints, exhibited large trial-to-trial spiking variability, and operated in an inhibition-stabilized regime. We derived analytical estimates on how training and network parameters constrain the changes in mean synaptic strength during training. Our results demonstrate that training recurrent neural networks subject to strong coupling constraints can result in connectivity structure and dynamic regime relevant to cortical circuits.</jats:p>