week 04/17

[ebridge2]

@ebridge2

• [x] for Thursday: test group-level qa with percent-retained (ie, retain highest 50% of edges, and threshold to 1 and remove below to 0), rank (replace value with the rank and use weighted network metrics) and squared-rank (ie, take rank, and square, to impose a more-sparse structure without totally removing small edges, again use weighted network metrics), as alternative to just thresholding by a absolute correlation value (what we are doing now). also, throw out really low things, but retain weights above.

• [x] a few subjects of kara's data do not work. investigate why; correct where possible. reg mc graph. See here for specific examples of one good, and the worst-performing, subjects. Relatively consistent performance on these brains.

• [x] add more qa statistics.

• [x] stretch: improve bids-crawl to work with non-complete sessions (ie, sessions that are missing anatomical, dwi, or functional data)

• [ ] Non FNGS goal: write presentation for thesis.

• [x] stretch: add slice timing correction to AWS batch deployment, and integrate tanay's cloud stuff/my adjustments to my working branch of ndmg. cloud

• [x] stretch: adapt existing R package to work with CPAC data. compare. Note that this is not populated, and the version of our pipeline is dated (from ~January, before we fixed nuisance and registration) but it will serve as a framework once we have AWS batch sorted out.

@ewalke31

• [x] for Thursday: begin working on preprocessing documentation https://neurodatadesign.github.io/fngs/ewalke31/tutorials/week_0410/Preprocessing.html
• [x] for Monday: do Motion Correction & Registration documentation https://neurodatadesign.github.io/fngs/ewalke31/tutorials/week_0410/Preprocessing.html https://neurodatadesign.github.io/fngs/ewalke31/tutorials/week_0410/Registration.html

@02agarwalt

• [x] implement AWS Batch functionality for FNGS pipeline: Link to relevant code (Proof of completion = demo)

[ebridge2]

• [ ] adjust bids-crawl to work with no session labels.

[jovo]

@ewalke31 documentation is looking great @02agarwalt i trust the demo went well? @ebridge2 nothing for me to look at?