Wider spread of excitatory neuron influence in association cortex than sensory cortex

Abstract

The basic structure of local cortical circuits, including the composition of cell types, is highly conserved across the cortical processing hierarchy. However, computational roles and the spatial and temporal properties of population activity differ fundamentally in sensory-level and association-level areas. In primary sensory cortex, the timescale of population activity is shorter and pairwise correlations decay more rapidly over distance between neurons, supporting a population code that is suited to encoding rapidly fluctuating sensory stimuli. In association cortex, the timescale of population activity is longer, and pairwise correlations are stronger over wider distances, a code that is suited to holding information in memory and driving behavior. Here, we tested whether these differences in population codes could potentially be explained by intrinsic differences in local network structure. We targeted single excitatory neurons optogenetically, while monitoring the surrounding ongoing population activity in sensory (auditory cortex) and association (posterior parietal cortex) areas in mice. While the temporal impacts of these perturbations were similar across regions, the spatial spread of excitatory influence was wider in association cortex than in sensory cortex. Our findings suggest that differences in recurrent connectivity could contribute to the different properties of population codes in sensory and association cortex, and imply that circuit models of cortical function should be tailored to the properties specific to individual regions. Significance statementCell-type-specific functional interactions and connectivity patterns have largely been studied in sensory cortex. Yet the properties of local network activity differ dramatically across the cortical hierarchy, possibly due to differences in intrinsic connectivity patterns. Here, we compared the functional impacts of individual excitatory neurons on local population activity, finding differences in the spatial spread of excitatory influence across regions. Our findings suggest that the structure of local networks differs across the cortical processing hierarchy, and these differences should be considered in circuit models of processes such as decision-making and working memory.