Description

Here, I will create QC images for each PLD. However, while going over the processing of multi-PLD, I come across a few issues, that I will list here. Comments from the closed duplicate #1209 added.

• Check individual TE_PLD_LD combinations: single-slice visualisation from all TE-PLD-LD combinations, taken from PWI3D &
  PWI3D.json
  -> in new folder with Multi-TE_PLD_LD_check
• Check TE & PLD fits: Create a NIfTI on the TE (&PLD?) fitting based on PWI3D, and exp(TE/T2*) and
  exp(PLD/T1b) -> same computation as in FABBER
• Think of quantitative values for QC, that issue a warning if outside a range?
• Create SD for all TE-PLD-LD combinations

A) single-PLD: here we have QC of mean PWI & temporal SD (tSD).
B) multi-PLD: we want to save the mean & tSD for each PLD separately. Also, we could do a separate simple ATT-fit.
C) multi-TE: we could save the mean (& tSD which would not be present for DEBBIE) for each TE. Also, we could do a separate simple TE-fit.
(see also #857 and #728)
Other QC ideas:

• an image showing a single slice for all PLDs, so we can check if the PLDs are all necessary and if we have a nice division of SNR from early to late phases.
• same for average values to detect scaling issues (as a plot over time)

A) We want SD for each TE/PLD/LD combination that has a significant number of volumes. E.g., if you have two PLDs with each 20 volumes, you want the SD for both. If you have 8 TEs with each 3 volumes, this might be an overkill. So we should think what we deem helpful.
B) We want to add other stats here, e.g., mean/SD/outlier mapping, maximal and minimal MIP might also give us insight over all TEs, fitting all TEs without any assumptions might also give us insight if things are wrong in the data or in the interaction of the data with a model.

Tasks

How to test

xASL_wrp_VisualQC_ASL

Release notes