Sensory systems are characterized by their ability to process large amount of input in parallel. The parallel processing often leads to multiple downstream pathways with diverse functionality. Some information pathways may be used to trigger rapid motor responses while others are wired for higher cognitive functions. How can we identify these functionally distinct pathways when the connectome for   a   nervous   system   is   available?     Here   we   propose   that   in   addition   to   conventional   adjacency matrices,   high order   connectivity   (indirect   connections   through   multiple   synapses)   should   also   be analyzed. This is based on the consideration that multi-synapse pathways may be more important, in terms of high level functionality than the direct or the shortest pathways. To test the power of our method,  we   analyzed   connectome   of  Caenorhabditis   elegans  and   found   that   the   high-order connectivity   of   the  C.   elegans  neural   network   is   associated   with   the   pathways   that   mediate   the complex social feeding behavior. We further applied our method to the analysis of a partial network of Drosophila central complex. Our analysis revealed the importance of a pair of atypical neurons in the high level signal processing. The exact function of the pair of neurons is still unknown, but their possible roles in maintaining integrity of the spatial representation will be discussed in the talk.