Speaker
Description
The highly recurrent networks of the cerebral cortex are thought to profoundly shape neural responses, amplifying certain patterns of activity over others. However, whether such selective amplification organizes network activity at the level of the millimeter-scale networks which underlie perception and behavior remains unknown. Here, we combine patterned optogenetic stimulation informed by a computational model with in vivo calcium imaging in immature ferret visual cortex to show that cortical networks are preferentially activated by inputs aligned to endogenous recurrent subnetworks. Motivated by our model indicating that spontaneous activity can provide a proxy for the millimeter-scale dominant functional subnetworks within recurrent cortical circuits, we found that the reliability and specificity of responses to input activity patterns was determined by the degree of overlap with the dominant modes of spontaneous activity. Inputs well matched to endogenous networks evoked reliable and specific responses that showed greater spatiotemporal stability than random misaligned input patterns, demonstrating the selective amplification of millimeter scale activity within cortical networks. This preferential amplification suggests that early cortical networks act as tuned filters, leveraging endogenous dynamics to stabilize and refine the precise sensory representations that emerge throughout development.