Dear Meishan,

When using the 'atlas file' option for a .nii ROI file in CONN which is simply a 0/1 mask, CONN will check if that mask is comprised of potentially multiple different non-contiguous anatomical areas or clusters (using SPM default 18-connectivity scheme to define contiguous areas; see "help spm_bwlabel" for details), and associate a separate ROI for each one of these clusters. 

Hope this helps

Alfonso

Originally posted by Meishan AI:

Dear Alfonso,

I imported an insula mask from an external resource. After I imported it to CONN and had the option "Atlas file" selected, there were two clusters appeared. However, when I was viewing it in fsleyes, it is simple a binary (with max=1 and min=0) of insula mask. 

In the second level analysis of CONN, I can see that there were normal time series signals in both clusters. I wonder if you have any insight in how CONN recognize different parcels in one file when importing ROIs and the potential interpretation of this binary file recognized as atlas file.

Thank you!

Meishan