| Popis: |
SummarySensory cortex transforms its inputs, generating neural responses that represent features of the sensory world. Dense local, recurrent connections are a major feature of cortical circuits, yet how they affect these transformations is unclear, with some studies reporting recurrent circuits amplify certain excitatory inputs, and other studies showing instead local suppression. Here we find that input to mouse V1 excitatory neurons generates salt-and-pepper patterns of excitation and suppression. We show that neurons’ responses are not strongly predicted by their optogenetic input alone, suggesting that recurrent inputs nonlinearly transform the population response. A balanced-state network model explains the observed dynamics, suppressed responses, and long tail of excited responses, but only with substantial variance in cell-to-cell connectivity. In sum, we find recurrent cortical connectivity has high variability and high average strength, which can decouple individual cells’ responses from the direct input each cell receives, leading to substantial suppression generated by excitatory input. |
