acquire and parse ajames data (T1+cowa+motor+rest)

[stowler]

R, 20140807:

Today @kmcgregor123456 xfered and unzipped Andy's data.

• It's nii, not dicom as Keith expected.
• keith will get slice timing info from Andy
• email to keith: How many runs should Stella have for each sequence? Spot checking, each participant in the archive seems to have one cowa run (150 TRs x 2 s), one rest run (225 TRs x 2 s), and one motor run (90 TRs x 2s).
• I aggregated the 3D geometry across participants to make it easier for Stella to identify inconsistencies

[stowler]

F, 20140808:

Keith/Stella/Andy call:

• Andy confirmed one run per paradigm is correct
• blind numbers are sequential in time
• switched from 8 to 32 ch coil between participants 49 and 52
• ...but that does not explain the 3x3x3 mm EPI voxels in 35, 36, 38, 39, 41 (Andy will check on cause)
• Keith and Bruce want us to QA and process data from both coils. There are pending decisions about whether both coils will make it into group analyses.
• participant 9 has the only non-oblique EPIs, but s/he is being excluded for non-image reasons anyway

[stowler]

Sun, 20140810:

• I added all of the acq information I have to the repo and described in README for reproducibility.
• emailed Andy/Keith/Stella: number of warm-up TRs?
• Andy replied: no warm-up TRs in niftis. He also offered to acquire example dicom data for us:

We acquired 3 dummy TRs (6 sec total) for all tasks. Philips acquires and discards these TRs; they are not part of the generated nifti files. Would it be helpful if I had our tech generate a dicom file for each of the three fMRI data acquisition settings? (8channel interleaved, 8channel default, 32channel sequential ascending)? We could run the scan for a single TR, just to generate a dicom from which to draw task information.

[stowler]

M, 20140811:

I replied yes to Andy's offer after consulting with Keith. (It gets us slice orientation and confirmation of the slice timing text files Andy sent.)

[stowler]

W, 20141105:

Emailed Andy asking re: old-coil sessions that have 3x3x3 mm voxels in EPIs. Participants 35, 36, 38, 39, 41

[stowler]

F, 20150220:

Andy replied to @smtran about the EPIs she is working with:

No, the files I sent are the “raw” files. They have not undergone any slice timing correction or stripe removal. However, I left out the subjects with the worst motion-related stripe artifact; I would argue that stripe removal is not necessary for the subjects I sent.

[stowler]

M, 20150223:

I asked Andy for his PROCESSED time series so that we can speed up future collaborations regardless of what Stella and her mentors decide about her current analysis. Andy transferred them to BITC, from which I'll transfer them to a read-only directory on hippostore after I confirm lack of VA-sensitive data.

Andy's description:

I initially offered to send the files post-processing: i.e. after despiking, slice correction, deobliquing, motion correction, co-registration with normalized (MNI) anatomic, regression of signal from CSF and WM, 6mm Gaussian FWHM spatial smoothing, quadratic detrending, scaling to 100% signal, and removal of stripes via ICA. That offer still stands, if you could use the post-processed data for your analyses. I would prefer that all data from the project undergo consistent post-processing, so that findings can be compared across sites; I am just not sure what steps you need (or don’t need) for your analyses.