Node Detection Using High-Dimensional Fuzzy Parcellation Applied to the Insular Cortex

摘要: Several functional connectivity approaches require the definition of a set regions interest (ROIs) that act as network nodes. Different methods have been developed to define these nodes and derive their effective connections, most which are rather complex. Here we aim propose relatively simple "one-step" border detection ROI estimation procedure employing fuzzy c-mean clustering algorithm. To test this explore insular beyond two/three-region model currently proposed in literature, parcellated cortex 20 healthy right-handed volunteers scanned resting state. By high-dimensional connectivity-based process, confirmed two patterns previously described. This method revealed complex pattern where detected clusters subdivided into several other networks, some not commonly associated with cortex, such default mode parts dorsal attentional network. Furthermore, was reliable, demonstrated by confirmative analysis performed on replication group subjects.