The effect of correlations in the background activity on the information transmission properties of neural populations

Recently, the information transmission properties of populations of neurons with independent noise inputs were examined and it was shown that noise can improve the transmission of sub-threshold signals. Information transmission is maximized at a certain noise level which, in general, depends on the population size. In the central nervous system of higher animals, however, the noise is likely to be correlated. In this paper we therefore investigate the effect of correlations between neurons on the information transmission properties of populations of neurons. We show that correlations in the noise inputs of neurons not only decrease information transmission but also immediately reduce the optimal population noise level to that of the single neuron. Hence, information about the population size does not need to be made available to the single neuron and therefore local adaptation rules as suggested in (Phys. Rev. Lett. 90 (2003) 120602) suffice.