[x] outline of FNGS overview paper. Deliverable: figure list. figure list
[x] rough draft of FNGS presentation.
[x] explore self-registration more in depth. Compare current procedure with EPI registration on extensive testing, and visualize lots of QC. this determination will be made with qualitative considerations. Deliverable: see below.
[x] test whatever version of FNGS is in the best condition on next Wednesday on a minimum of BNU1, BNU2, SWU4, DC1, HNU1, LMU1, NKI1, and any other datasets greg wants. Deliverable: ndmg plots (@gkiar let me know which you think are valuable to reproduce from ndmg draft, just so we can see them for reference to decide whether we want to add to ndmg paper). add all code in an intuitive-to-call way to Discriminability repo repo
[x] joint embedding. Deliverable: one of the joint embedding tricks from previous deliverable completed.
ADDED 3/24
[x] PIVOT from JE: develop new statistic for assessing registration quality, and evaluate with a simplified algorithms.md (note that the majority of the functionality is bet hyperparameter searching and thresholding, which is implemented by FSL already). registrationspecs
[x] "robust registration", rewrite all QC to simplify/audit, Deliverable: new pull request. PR and +1200 lines of python
[x] Re-run MNI-EPI-Template + FLIRT pipeline on BNU1 and LMU1
[x] Create MNI-EPI-Template + FLIRT + FNIRT pipelines using different FNIRT hyperparameters. Run on BNU1 (not on LMU1 because of space and time restrictions)
@ebridge2
ADDED 3/24
Thesis:
@02agarwalt
@ewalke31