Jorge F. Mejías
SIGNAL DETECTION IN NEURAL POPULATIONS: THE IMPORTANCE OF HETEROGENEITY
It is known that neural systems display a prominent heterogeneity in individual neuron properties, even among populations of same-class neurons. However, the effect of such heterogeneity in the dynamics of neural populations has not been fully understood up to date. Here, we present a detailed theoretical and numerical study of the implications of
heterogeneity in the properties of neural populations. Our study reveals important effects of heterogeneity in several properties of interest, such as the mean firing rate, the irregularity of the mean activity, or the synchronization properties of the network. We also study the role of heterogeneity in networks with realistic synaptic dynamics (such as short-term depression). Finally, our results show that heterogeneity in neural excitability allows for different information processing strategies, where the concrete strategy used depends on the level of correlation among individual neurons. This establishes a direct relation between (optimal) heterogeneity levels and neural correlation levels.