Item

Abstract

We present an anatomically constrained model of the dense wiring diagram of thalamoand intra-cortical circuits in rat barrel cortex. The model is based on the assumptions that synaptic contacts are formed (1) randomly { i.e. proportional to the locally available of pre- and postsynaptic target structures { and (2) independent of each other. These two assumptions can be regarded as a 'null-hypothesis' for non-specificity in cortical connectivity (commonly referred to as Peters' Rule), allowing to test whether wiring diagrams as predicted from these assumptions are in line with different experimental measurements. We show that the model reproduces measurements of 1st order statistics (e.g. pair-wise connection probability) for all cell types and layers, and that it predicts highly non-random 2nd and higher order connectivity patterns. We find that non-random (i.e. specific) connectivity patterns in cortex do not necessarily originate from local specificity rules, but reflect cell type- and location-specific organizational principles at population, cellular and morphological levels.