| The synaptic circuits of the cerebral cortex are still poorly understood, yet knowing their basic structure appears key for understanding their function (Lorente de No, 1949). While some argue that there is a basic modular circuit present in all cortical regions (Douglas et al., 1989; Hubel and Wiesel, 1977), others suggest that synaptic circuits could be randomly structured (Braitenberg and Schuz, 1998). To investigate the patterns of synaptic connections present in neocortex, I have developed a novel two-photon optical mapping method (Nikolenko et al., 2007) to systematically reveal cells that connect to four classes of neurons in slices of mouse primary sensory cortex. Inputs to these cells originated preferentially from specific cortical layers and often were laterally restricted, revealing functional columnar circuits with sharp boundaries. Moreover, many neurons extensively sampled particular territories, and, in some cases, virtually every cell from a particular layer was connected to the postsynaptic target. The results reveal circuits with dense columnar connectivity, approximating in some cases the complete sampling from every potential presynaptic cell in an input layer. I discuss the implications of these findings in the context of the computational strategies used by the cortex. |
