Calculate regressors and save them in _confounds.tsv table

[oesteban]

• confounds in native space
  • Frame Displacement regressor (input is HMC movpar)
  • after HMC & unwarping:
    • White Matter, Grey Matter, CSF signal regressors, Global Signal regressor
    • dvars
    • TCompCor
    • ACompCor
• Put outputs in bids-derivatives according to standard
• integration test
• headers in confound file/derivative

[berleant]

I have the same concern as in #44

[oesteban]

• For CompCor, please see New CompCor interface nipy/nipype#1594
• For the Global Signal and the tissue-wise signal regressors we could submit another new interface to nipype.

OR

• Have only one interface for all of this (ComCor, Global Signal, tissue-wise signal regressors).

In both cases, the FD regressor is a bit more complicated and we already have all the building blocks in place to generate it. That's why I would not mix it with the rest of them.

I prefer the two separated interfaces, and let it up to the fmriprep to merge all of them in a single file (using AddCSVColumn, for instance)

WDYT @chrisfilo ?

[berleant]

@oesteban At your suggestion, I took a look at AddCSVColumn in nipype.algorithms.misc, and it doesn't match the requirements for this project as written--would you recommend I modify/refactor AddCSVColumn to make it work (may take longer, but perhaps more useful in the future), or write a simple T/CSV manipulation function just for this (less work now, but not as useful to others)?

[oesteban]

@shoshber @chrisfilo let's go for a quick and dirt implementation, probably just with something like numpy.savetxt(delimiter='\t')

[chrisgorgo]

Agreed - or a Function Interface that uses pandas.

On Fri, Sep 30, 2016 at 2:03 PM, Oscar Esteban notifications@github.com
wrote:

@shoshber https://github.com/shoshber @chrisfilo
https://github.com/chrisfilo let's go for a quick and dirt
implementation, probably just with something like
numpy.savetxt(delimiter='\t')

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#109 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAOkp9KlMMUHRP39KgSxqZZgnOpE5gdWks5qvXktgaJpZM4JkGFS
.