Hi Alex,

The standard way to do this is by performing seed-based connectivity analyses (computing connectivity between a seed ROI and each voxel in the brain), e.g. running SBC first-level analyses in CONN.

That said, this is not exactly the same as what IC or GCOR would translate to if, instead of the entire brain, we restrict the target voxels to an ROI mask. The main reason is because in SBC analyses the "averaging" is happening before computing correlations (i.e. the entire ROI is represented by its average BOLD timeseries, and then the correlation between that ROI-level BOLD timeseries and each individual voxel in the brain is computed), while in GCOR or IC the "averaging" is happening after computing correlations (i.e. the correlations between a voxel and the rest of the brain are computed separately for each target voxel, and then averaged across all voxels in the brain -or across all voxels within an ROI if I am interpreting correctly what you are proposing-). If you want to perform the latter, one way to do that is to use the function conn_vv2rr (this assumes you have voxel-to-voxel analyses enabled in your project). You can, for example, using the syntax:

datafile = conn_vv2rr('myROIfile.nii', 'style','vv2rv');

will create a data file containing the average correlation between each voxel in the brain and all voxels within your ROI (ROI specified in the "myROIfile.nii" file). Let me know if that is what you had in mind (if so see "help conn_vv2rr" for additional options/details) and/or if I am misinterpreting. 

(and regarding your second question, yes, the 'import values' option after using GCOR or IC would just average the computed network centrality measures across all voxels within an ROI)

Best
Alfonso
Originally posted by Alexandra Muratore: