SimNet: A new algorithm for measuring brain networks similarity
Résumé
A very critical issue in many applications is how to compute graphs similarity. Many algorithms have been proposed to calculate similarity between graphs. Most of them ignore the physical location of nodes which is a key factor in the context of brain networks involving spatially defined functional areas. The main purpose of this abstract is to present a new algorithm called "SimNet" to measure the similarity between tow graphs. The main originality of the algorithm is taking into account the spatial location of the graphs, in addition the nodes and the edges similarity. The performance of the proposed algorithm is analyzed on synthetic graphs. The algorithm is then evaluated on a real brain networks obtained during a picture naming task and shows a good performance to detect modules of brain networks that share the same spatial locations.