Hi,

Here's a question that I raised a fair while ago (JUNE 2017) and Florian answered by email; I'm posting to make it generally accessible. His response was helpful, so I'm appending it, too.  I apologize for my delay, I'm a newbie, and an old-bie (soon facing my Beatles bday "When I'm 64").  Forums are new to me, too. But here it is...

I wanted to back reconstruct from one manually selected IC to the whole brain. 

"I ran a single-subject brainstem-masked mICA and the results are in a folder for "subj01".

"So I'm guessing the new "input" is original whole brain functional? (from a FEAT directory, in this case). I left the mask entry open, although I imagine there's a way to use the selected IC here?? Not sure how to do that.

"I chose Inversion.I entered an 'input_preproc.txt' file. When I say GO, it says "Please select ICs compnents directory to perform back reconstruction." So, I set the "dim0" folder as the ICA directory.

"Then it says "Custom mask is either empty or the mask file was not found."
But I want to back-reconstruct to the whole brain, so why is a mask needed?"

(end of initial question)

-- Here's Florian's reply:

"In the current version of the toolbox you have to select a mask. In your case, this should be a whole-brain mask matching your whole-brain functionals. Such a mask should have been automatically generated by FEAT. In a standard preprocessing it is simply called "mask.nii.gz" and sits in the feat folder."

Thanks Florian!!

Ron