Understanding the genesis of shared trial-to-trial variability in neural activity within sensory cortex is critical to uncovering the biological basis of information processing in the brain. Shared variability is often a reflection of the structure of cortical connectivity since this variability likely arises, in part, from local circuit inputs. A series of experiments from segregated networks of (excitatory) pyramidal neurons in mouse primary visual cortex challenge this view. Specifically, the across-network correlations were found to be larger than predicted given the known weak cross-network connectivity. We aim to uncover the circuit mechanisms responsible for these enhanced correlations through biologically motivated cortical circuit models. Our central finding is that coupling each excitatory subpopulation with a specific inhibitory subpopulation provides the most robust network-intrinsic solution in shaping these enhanced correlations. This result argues for the existence of excitatory-inhibitory functional assemblies in early sensory areas which mirror not just response properties but also connectivity between pyramidal cells.

Cortical state is modulated by myriad cognitive and physiological mechanisms. Yet it is still unclear how changes in cortical state relate to changes in neuronal processing. Previous studies have reported state dependent changes in response gain or population-wide shared variability, motivated by the fact that both are important determinants of the performance of any population code. However, if the state-conditioned cortical regime is well-captured by a linear input-output response (as is often the case), then the linear Fisher information (FI) about a stimulus available to a decoder is invariant to state changes. In this study we show that by contrast, when one restricts a decoder to a subset of a cortical population, information within the subpopulation can increase through a modulation of cortical state. A clear example of such a subpopulation code is one in which decoders only receive projections from excitatory cells in a recurrent excitatory/inhibitory (E/I) network. We demonstrate the counterintuitive fact that when decoding only from E cells, it is exclusively the I cell response gain and connectivity which govern how information changes. Additionally, we propose a parametrically simplified approach to studying the effect of state change on subpopulation codes. Our results reveal the importance of inhibitory circuitry in modulating information flow in recurrent cortical networks, and establish a framework in which to develop deeper mechanistic insight into the impact of cortical state changes on information processing in these circuits.