ABSTRACT
A cortical network is a complex system that contains multiple interacting circuits, generating multiple types of lateral interactions among cortical columns. Minicolumns are 0.05 mm diameter cords of cells extending radially across all cortical layers, whose receptive field and functional properties distinguish each such minicolumn from its immediate neighbors. The stimulus tuning properties of the cells in a cortical network set limits on which statistical relations present in the environment can be captured by the lateral connections. While lateral connections are the core of the cortical network in its role as a statistical device, what makes this device functionally useful and gives it its inferential power are two other putative properties of the cortical network. Long-range interactions are believed to be responsible for integration of sensory information coming to the cortical network from different parts of spatially extensive stimulus patterns.
