Hello,

I am using the  "mask with grey matter" option and I'm curious if this is using the gray matter from the structural or functional dataset. I used high-resolution imaging  during acquisition several years ago with a restricted FoV  to optimize imaging of the MTL. I'd like to us the gray matter masking feature to not extract from some of the superior regions (SFG, pre-central gyrus) where there is no valid data. But this would require masking with the functional gray matter segementation.

Thank you!
Sarah

Hi Sarah,

The "mask with grey matter" option always uses the "gray matter" ROI as defined in the same project (in Setup.ROIs). Typically (e.g. if you have used CONN's default preprocessing pipeline) those ROIs are automatically populated during the structural segmentation step as part of preprocessing, so they will be derived from the segmentation of the structural data. If you prefer to use masks derived from the segmentation of the functional data instead, you may, for example, simply switch the relative order of the structural and functional normalization/segmentation steps in CONN's default preprocessing pipeline (whichever step happens last will be the one that populates the gray/white/csf ROIs in your project), or manually select your desired gray matter ROI masks in Setup.ROIs.

Hope this helps
Alfonso

Originally posted by Sarah Kark:

Dear Alfonso,

Thank you for your help and explanation.

My goal is to obtain REX ROI activity extractions for atlas ROIs, but to not draw on voxels activity from outside of the field of view (because of course those voxels would not be valid). See top left image in the attached for a group-level mask I made from threshold the wau* files (these data will not be smoothed). These are voxels that the group have in common and show the limited FOV.

I have been thinking of several ways to try to accomplish my goal. 1) Enter this group mask back into Setup.options in place of the default Analysis mask (mask.volume) to restrict all analyses to the voxels in the FoV or 2) As you suggested, change the Setup.ROI Grey Matter ROIs to the functional gray matter files (which are the wc1art_mean_au*.nii files?) and then pressing DONE and re-running the de-noising stage. Does either approach work?

Finally, in a first pass of this analysis, errors would be thrown for ROI #59 for REX extraction step that "warning: no valid data for" ROI #59, yet in the ROI resulting files in the data structure there were still TR-wise numbers for those ROIs. Is the data were not valid, what is being reported there? Should those numbers be NaNs? it was correct for ROI #59 to throw an error because it is the precentral gyrus and is outside of the FOV, but there are others that also should have had this error.

Thank you!
Sarah Kark

I suppose it would change the Grey Matter, White Matter, and CSF Setup.ROIs to the functional ones (wc1art_mean_au, wc2art_mean_au, and wc3art_mean_au) so they all reflect the restricted FOV in the functional data for aCompCorr, is that right?