Spontaneous and stimulus-induced coherent states of dynamically balanced neuronal networks

How the information microscopically processed by individual neurons is integrated and used in organising the macroscopic behaviour of an animal is a central question in neuroscience. Coherence of dynamics over different scales has been suggested as a clue to the mechanisms underlying this integration. Balanced excitation and inhibition amplify microscopic fluctuations to a macroscopic level and may provide a mechanism for generating coherent dynamics over the two scales. Previous theories of brain dynamics, however, have been restricted to cases in which population-averaged activities have been constrained to constant values, that is, to cases with no macroscopic degrees of freedom. In the present study, we investigate balanced neuronal networks with a nonzero number of macroscopic degrees of freedom coupled to microscopic degrees of freedom. In these networks, amplified microscopic fluctuations drive the macroscopic dynamics, while the macroscopic dynamics determine the statistics of the microscopic fluctuations. We develop a novel type of mean-field theory applicable to this class of interscale interactions, for which an analytical approach has previously been unknown. Irregular macroscopic rhythms similar to those observed in the brain emerge spontaneously as a result of such interactions. Microscopic inputs to a small number of neurons effectively entrain the whole network through the amplification mechanism. Neuronal responses become coherent as the magnitude of either the balanced excitation and inhibition or the external inputs is increased. Our mean-field theory successfully predicts the behaviour of the model. Our numerical results further suggest that the coherent dynamics can be used for selective read-out of information. In conclusion, our results show a novel form of neuronal information processing that bridges different scales, and advance our understanding of the brain.