search

PennLINC / xcpEngine

Official public repository for the XCP Engine. This tool is deprecated in favor of XCP-D and ASLPrep.
MIT License
66 stars 42 forks source link

Number of PCA components in acompcor #388

Open YukiSakai1209 opened 4 years ago

YukiSakai1209 commented 4 years ago

Dear xcpEngine team, thanks for putting such an excellent pipeline.

Concerning acompcor, I wonder how many components will be considered for covariates? ⁃ In xcpengine:1.0, according to outputs, six components of acompcor seems to be considered. ⁃ In xcpengine:1.2.3, five components.

In web documentation (https://xcpengine.readthedocs.io/config/streams/fc.html?highlight=acompcor#functional-connectivity-streams), a sufficient number of signals to explain 50 percent of the variance in white matter and cerebrospinal fluid were included. So, it should depend on each data. However, according to https://github.com/PennBBL/xcpEngine/issues/334, is it fixed to the first five components?? I’m confused about the way to determine the number of PCA components of acompcor.

a3sha2 commented 4 years ago

hi @YukiSakai1209 which frmripep version did you run?

YukiSakai1209 commented 4 years ago

Thank you for your quick response! I used fmriprep 1.3.2.