migrate off-line OS X processing to WMB ubuntu 14.04

[stowler]

I'm adapting each of these steps for the ubuntu workstations in WMB (pano/rama), and testing to ensure that single-session and batch processing works for current and future acquisitions:

• [x] renaming/scrubbing session-level directories (cda001_YYYYMMDD -> cda001pre)
• [x] homogenization of series-level naming (anat, fmri.rest.run1, etc.)
• [x] conversion of DICOMs to NIFTIs
• [x] skull-strip anatomic T1 based on manually-defined center-of-gravity
• [x] FBIRN quality control for fmri task
• [x] FBIRN quality control for fmri rest
• [x] generation of melodic .fsf files for fmri task
• [x] generation of melodic .fsf files for fmri rest
• [x] execution of initial melodic ICA for fmri task
• [x] execution of initial melodic ICA for fmri rest
• [x] initial de-noising of fmri task with FSL FIX (Standard20)
• [x] initial de-noising of fmri rest with FSL FIX (Standard20)
• [ ] hand-hold atypical acquisitions on a way that creates reproducible but non-sensitive script

Alternative de-noising of taskCMG using range of FIX parameters (trained-weights data and FIX threshold):

• [x] FIX trained-weights "Standard", FIX threshold 18
• [x] FIX trained-weights "Standard", FIX threshold 15
• [x] FIX trained-weights "Standard", FIX threshold 10
• [x] FIX trained-weights "WhII_Standard", FIX threshold 20
• [x] FIX trained-weights "WhII_Standard", FIX threshold 18
• [x] FIX trained-weights "WhII_Standard", FIX threshold 15
• [x] FIX trained-weights "WhII_Standard", FIX threshold 10

Alternative de-noising of resting state using range of FIX parameters (trained-weights data and FIX threshold):

• [x] FIX trained-weights "Standard", FIX threshold 18
• [ ] FIX trained-weights "Standard", FIX threshold 15
• [ ] FIX trained-weights "Standard", FIX threshold 10
• [ ] FIX trained-weights "WhII_Standard", FIX threshold 20
• [ ] FIX trained-weights "WhII_Standard", FIX threshold 18
• [ ] FIX trained-weights "WhII_Standard", FIX threshold 15
• [ ] FIX trained-weights "WhII_Standard", FIX threshold 10

Additional work to complete on pano/rama before Dec 1:

• [x] move skull-strip parameters and centers-of-gravity out of image dirs and into git repo
• [x] Monti lab's "optiBet" optimized skull-strip plus nonlinear registration
• [ ] initial group-level FMRI analyses
• [ ] freesurfer
• [ ] ICA-AROMA (maybe)

[stowler]

Emailed Madeleine and Joe:

Hi guys,

If you meet with me in person today we'll talk about this, but just so you have the links in your email here's an easy way for you to track the migration of your processing to WMB:

Madeleine: https://github.com/CVNRneuroimaging/proj.mh.cda2/issues/1

Joe: https://github.com/CVNRneuroimaging/proj.jn.cda2/issues/1

Once all of those items are checked-off and there are no "TBD's" left it will mean that all of your processing has moved to WMB, and you have complete single-session data sets that were built:

• on a consistent platform (i.e., everything from ubuntu 14.04),
• by consistent versions of FSL (i.e., no version variability across participants),
• with consistent scripts (the version of the scripts in your github repositories).

The end result of that is your data are reproducible by anyone, extendable to new acquisitions, and free of between-session/between-participant variability introduced by differences in software versions.

[stowler]

Launched execution of initial melodic ICA for fmri rest. (After omitting sessions with atypical task timing).

[stowler]

Launched FIX for resting state: melodicFixStandard20-gnuParallel15.ica

[stowler]

Launched FIX for taskCMG: melodicFixStandard18-gnuParallel15.ica

[stowler]

Launched final rama -> pano backup in preparation for devoting rama to Bruce's data.

[stowler]

Launched FIX for taskCMG: melodicFixStandard15-gnuParallel15.ica

[stowler]

Initial testing of optiBET.sh (cda006pre).

Processing took about 26 minutes on tasaki.

Inspected results with bwFslviewGreyOnGreen.sh. Shockingly good.

TBD: next tests are cda005, cda008, cda104 (Simone’s difficult extractions) TBD: resolve pano’s getopts incompatibility TBD: maybe edit optiBET.sh to allow center-of-gravity as command-line argument.

[stowler]

Launched FIX for taskCMG: melodicFixStandard10-gnuParallel12.ica

[stowler]

Launched optiBET parallel-sessions batch

[stowler]

Launched FIX for RSFMRI: melodicFixStandard18-gnuParallel10.ica

[stowler]

Launched FIX for taskCMG: melodicFixWhII_Standard20-gnuParallel10.ica

[stowler]

Launched FIX for taskCMG: melodicFixWhII_Standard18-gnuParallel10.ica

[stowler]

Launched FIX for taskCMG: melodicFixWhII_Standard15-gnuParallel10.ica ...operator error: I accidentally launched 08.3.fix-rest-parallelRuns.sh instead