Impact of Excitation-Inhibition Balance/Imbalance on Dynamics of Cortical Neural Networks

The purpose of this research is to study the implications of Excitation/Inhibition balance and imbalance on the dynamics of ongoing (spontaneous) neural activity in the cerebral cortex region of the brain. The first research work addresses the question that why among the continuum of Excitation-Inhibition balance configurations, particular configuration should be favored? We calculate the entropy of neural network dynamics by studying an analytically tractable network of binary neurons. Our main result from this work is that the entropy maximizes at regime which is neither excitation-dominant nor inhibition-dominant but at the boundary of both. Along this boundary we see there is a trade-off between high and robust entropy. Weak synapse strengths yield entropy which is high but drops rapidly under parameter change. Strong synapse strengths, on the other hand yield a lower, but more robust, network entropy. The second research work is motivated from experiments suggest that the cerebral cortex can also operate near a critical phase transition. It has been observed in many physical systems that the governing physical laws obey a fractal symmetry near critical phase transition. This symmetry exists irrespective of the observational length-scale. Thus, we hypothesize that the laws governing cortical dynamics may obey scale-change symmetry. We test and confirm this hypothesis using two different computational models. Further, we extend the transformational scheme show that as a mouse awakens from anesthesia, scale-change symmetry emerges. The third research project is motivated by experimental observations from in motor cortex under modulation of inhibitory inputs. We found that low intensity increase (decrease) in overall inhibition in cortex causes decrease (increase) in spiking activity for some neurons. Even though, the population level activity largely unchanged. This behavior is paradoxical when compared to the status quo that says that increase (decrease) inhibition should lead to decrease (increase) in neural spiking activity. We simulated similar dynamical change to inhibitory signal modulation in neural network model. We found that this paradoxical behavior arises due to sparse connectivity and inhomogeneity in inhibitory weights