Title: Effect in the spectra of eigenvalues and dynamics of RNNs trained with Excitatory-Inhibitory constraint

Abstract: In order to comprehend and enhance models that describe various brain regions is important to study the dynamics of trained recurrent neural networks. Including Dale\textquotesingle s law in such models usually presents several challenges. However, this is an important aspect that allows computational models to better capture the characteristics of the brain. Here we present a framework that enables networks to be trained using such a constraint. We have used it to train them in simple decision-making tasks. We characterized the eigenvalue distributions of the recurrent weight matrices of such networks. Interestingly, we discovered that the non-dominant eigenvalues of the recurrent weight matrix are distributed in a circle with a radius less than 1 for those whose initial condition before training was random normal and in a ring for those whose initial condition was random orthogonal. In both cases, the radius does not depend on the fraction of excitatory and inhibitory units nor the size of the network. Diminution of the radius, compared to networks trained without the constraint, has implications on the activity and dynamics that we discussed here.