Hello members, 

  I use CONN to look at the resting-state functional connectivity in young and elderly. My data was preprocessed using spm8 DARTEL. The GM, WM and CSF masks were generated from segmentation. Only the GM mask was later normalised and smoothed, but not for WM and CSF. If the WM and CSF masks are left blank in CONN, segmentation for WM and CSF will be done from the structural images (normalized). I'd like to know if the CONN generated WM and CSF masks are normalised by default? That would mean the masks would be normalised twice in my case? It would be really helpful to know what is recommended to input to CONN if the normalized WM and CSF masks are not available from the DARTEL preprocessing pipeline? 

  If I want to look at the functional connectivity between two functional networks, each of my ROI.nii has multiple regions (attached an example), the graphical presentation of the results in the second level analysis display one spot for each network. I'd like to know how CONN deals with ROI images with multiple regions? And what the single spot display in the second level analysis represents? 

Many thanks in advance. 

Best wishes,

Belle