Small world coefficient

In Graphvar you will have to check the box for computing smallworldness. It would be important to make sure that you generate sufficient random networks per subject (i.e. 100-1000) to obtain a reliable measure of smallworldness (smallworldness is the ratio of the clustering coeff and the pathlength as compared to random networks). Best, Johann

Hey, You could read some paper about small-worldness: Latora V, Marchiori M. (2003): Economic small-world behavior in weighted networks. The European Physical Journal B - Condensed Matter and Complex Systems V32(2):249-263. Achard S, Salvador R, Whitcher B, Suckling J, Bullmore E. (2006): A resilient, low-frequency, small-world human brain functional network with highly connected association cortical hubs. J Neurosci 26(1):63-72. He Y, Chen ZJ, Evans AC. (2007): Small-world anatomical networks in the human brain revealed by cortical thickness from MRI. Cereb Cortex 17(10):2407-19. The small-worldness included the coefficient as following: Cp: clustering coefficient of network; Lp: shortest path length of network; Gamma: the ratio of Cp and mean value of Cprand (random networks' Cp), Cp/mean(Cprand); Lambda: the ratio of Lp and mean value of Lprand (random networks' Lp), Lp/mean(Lprand); Sigma: Gamma/Lambda. And the AUC of (area under curve of all threshold) these coefficient for fMRI network. As Dr. Kruschwitz said, 100 random should be necessary as usual. You could use GraphVar or GRETNA to do it. For GraphVar, please read Dr. Kruschwitz's reply. For GRETNA, you need import your raw network matrix (*.txt or *.mat), it will help you construct network based on threshold tyep and range. Download and read our manual for the details. http://www.nitrc.org/projects/gretna/ Best, Sandy