search

PennLINC / xcpEngine

Official public repository for the XCP Engine. This tool is deprecated in favor of XCP-D and ASLPrep.
MIT License
66 stars 42 forks source link

Invert tmask to interpolate the correct timepoints #474

Open dhasegan opened 2 years ago

dhasegan commented 2 years ago

Discussion is here: https://github.com/PennLINC/xcpEngine/issues/473

I am still debugging this so don't pull this in yet.

dhasegan commented 2 years ago

do I have to change tmask to ones in utils/tfilter ?

tmask=ones

to:

tmask=zeros
dhasegan commented 2 years ago

I tested the last version and it correctly processed my volumes (both censored and not). I think it's safe to push this and I can chat about it if needed.

dhasegan commented 2 years ago

For example, my qcfc figure resulting from 125 volumes censored (using a variation of the fc-36p_scrub.dsn):

Screen Shot 2022-01-12 at 10 42 18 AM
a3sha2 commented 2 years ago

that figure is correct for scrubbing outputs. It is necessary to indicate which volumes were flagged, that is why you see c two residual images in regress folder, one without scrubbed volumes and the other with scrubbed volumes

dhasegan commented 2 years ago

I see them, its the _residualized_dtseries.svg and _residualized_hemi-L_bold.func.svg for left hemisphere.

Looks good to me. Is there anything else that I should do for this PR?