Notes on Visual Cortical Feedback and Feedforward Connections.

Abstract

The largely dissociated laminar distribution patterns of feedforward (FFD) and feedback (FBK) cortical connnections in sensory areas of nonhuman primates (NHP) have figured prominently in ideas of network organization and cortical hierarchy. While convenient, this emphasis does not take full account of the actual anatomical complexity. There is a diversity of FFD and FBK neuronal subtypes (cf. the corticogeniculate (CG) projections, where the fact of multiple CG neuronal subtypes inspired ideas of parallel processing). There is a complex architecture of recurrent connections, where FFD and FBK connections interact together and with numerous other cortical and subcortical connections, in ways still to be elucidated. Importantly, the laminar patterns so apparent at the global level, are less so from the perspective of postsynaptic dendritic input maps. Both supragranular FFD and FBK projecting neurons have apical dendrites in layer 1, and both FFD and FBK neurons can directly access inputs to layer 4 (respectively, via basal dendrites of neurons in deep layer 3 and apical denrites of infragranular neurons). Here, I briefly discuss some of the basic anatomical questions regarding FFD- FBK processing, still unanswered at the cellular level, and propose that the corrrelation with a functional hierarchy is less evident than often suggested.