I would like to use conn_batch to preprocess my data with a custom preprocessing pipeline.  This custom pipeline would do the following

1) Run ART
2) Smooth the functional data.

Is this possible?  The issue I am encountering that I haven't figured out a work around is specifying the motion correction parameters that ART needs through the conn_batch structure. 

Any suggestions you have would be greatly appreciated.

Thanks

Bob

Hi Bob,

When running ART from CONN there are currently two alternative ways to specify the subject-motion parameter file:

a) enter the motion parameter first into CONN as a first-level covariate named 'realignment' (if a first-level covariate with that name exists ART will use for each subject/session the motion file entered in this covariate for the same subject/session)
or b) otherwise, have the motion parameter file to be in the same folder as the functional data and named rp_{root}.txt (e.g. rp_file.txt), with your functional files named {prefixes}{root}.nii (e.g. swaufile.nii; if using multiple 3d volumes instead, the first functional volume is expected to be named {prefixes}{root}.img, e.g. motion file rp_file001.txt associated with the first functional file swaufile001.img)

In your case I would probably suggest to:

1) first use batch.Setup to enter your almost-fully-preprocessed data into CONN, including your functional data as well as your realignment parameters as a 'realignment' first-level covariate (run this Setup batch but do not set the batch.Setup.done field to 1 just jet)

2) then run the additional preprocessing steps adding the following line to your script:

  conn_setup_preproc({'functional_art','functional_smooth'},'art_thresholds',[9 2],'fwhm',8);

3) and then finish the Setup step using batch.Setup.done=1 and continue normally from there

Let me know if you would like me to further clarify any of the above
Best
Alfonso 
Originally posted by Bob Kraft: