Dear Melodic experts,
I'm analyzing resting fMRI data using Melodic and dual regression. How can I identify in FSL certain network such as the salience network, which is not included in templates such as Beckmann's 2005 Royal Society paper and Smith's 2009 PNAS paper?
Is the following application of the goodness of fit procedure taken by Seeley et al, 2007 in Journal of Neuroscience acceptable?
1. Creating of the salience network template from atlas, which comprises e.g. the ACC and insula.
2. Calculation of goodness of fit for each IC: the mean of z-score of voxels falling within the template minus the mean of z-score of voxels outside the template
3. Selecting the IC as the salience network which has the largest goodness of fit value.
Thanks in advance,
Jun Miyata
|