Skip to content

add DVFinal CPAC bug to divergences notes #329

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
nx10 merged 2 commits into from
May 9, 2026
Merged

add DVFinal CPAC bug to divergences notes #329

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8573ec0
add DVFinal difference
jpillai00 May 7, 2026
cc414a0
add cpac path to xcp file
jpillai00 May 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion tests/data/cpac_rbc_divergences.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
Findings from cross-referencing C-PAC source code and actual RBC pipeline run logs.

**Sources:**
- C-PAC code: `CPAC/anat_preproc/anat_preproc.py`, `CPAC/anat_preproc/ants.py`, `CPAC/pipeline/cpac_pipeline.py`, `CPAC/func_preproc/func_preproc.py`, `CPAC/alff/alff.py`
- C-PAC code: `CPAC/anat_preproc/anat_preproc.py`, `CPAC/anat_preproc/ants.py`, `CPAC/pipeline/cpac_pipeline.py`, `CPAC/func_preproc/func_preproc.py`, `CPAC/alff/alff.py`, `CPAC/qc/xcp.py`
- C-PAC RBC run logs (ds000001, sub-01_ses-1)

---
Expand Down Expand Up @@ -62,6 +62,16 @@ This bug affects all outputs downstream of the frequency filter: cleaned BOLD, R

---

## 5. DVFinal metrics computed from pre-regression BOLD (same as DVInit)

The XCP-style QC file reports `meanDVFinal` and `motionDVCorrFinal` as post-regression DVARS metrics (per the XCP dictionary: "correlation of RMS and DVARS after regression"). However, inspecting the C-PAC source code (`xcp.py`) reveals that both `DVInit` and `DVFinal` are computed from the same pre-nuisance regression BOLD input.

Comment thread
jpillai00 marked this conversation as resolved.
This was confirmed empirically: `meanDVInit = meanDVFinal` and `motionDVCorrInit = motionDVCorrFinal` for every subject across all three datasets (HBN, PNC, NKI) during the small benchmarking study.

In contrast, RBC correctly computes `DVFinal` from the post-regression bandpass-filtered BOLD (`cleaned_bold`), as intended.

---

## Notes

### Two parallel motion correction paths
Expand Down
Loading