Skip to content

ENH: One step interpolation for CIFTI subcortical structures - #558

Open
mgxd wants to merge 8 commits into
masterfrom
enh/subcortical-mni6
Open

ENH: One step interpolation for CIFTI subcortical structures#558
mgxd wants to merge 8 commits into
masterfrom
enh/subcortical-mni6

Conversation

@mgxd

@mgxd mgxd commented Jun 26, 2026

Copy link
Copy Markdown
Collaborator

Resample BOLD directly into MNI152NLin6Asym for the CIFTI subcortical grayordinates, rather than native -> MNIInfant -> per roi MNI152NLin6Asym realignment, and add a QC figure.

Formerly, we resampled BOLD to MNIInfant, then ran a separate per-structure MNIInfant -> MNI152NLin6Asym refinement to reach the HCP grayordinate space. The anatomical workflow already produces a direct anat -> MNI152NLin6Asym transform (optionally with multi-step registration), so BOLD can resample straight to MNI152NLin6Asym and label subcortical grayordinates from the HCP dseg. Initial comparisons show improvement, but additional testing is needed.

Immediate improvements:

  • single interpolation: timeseries is only interpolated once (native -> MNI152NLin6Asym), reducing spatial blurring and added temporal autocorrelation.
  • less compute, memory, and disk: drops the intermediate MNIInfant BOLD resample and the per-ROI registration/resample/smooth stage (e.g. the FLIRT node alone reserved ~4 GB), cutting nodes, working-directory footprint, and peak memory.

@codecov

codecov Bot commented Jun 26, 2026

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 50.79365% with 31 lines in your changes missing coverage. Please review.
✅ Project coverage is 62.22%. Comparing base (a2f98a0) to head (02f714e).
⚠️ Report is 7 commits behind head on master.

Files with missing lines Patch % Lines
nibabies/interfaces/reports.py 27.90% 31 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##           master     #558      +/-   ##
==========================================
- Coverage   62.69%   62.22%   -0.47%     
==========================================
  Files          74       73       -1     
  Lines        7237     7212      -25     
  Branches      891      887       -4     
==========================================
- Hits         4537     4488      -49     
- Misses       2428     2455      +27     
+ Partials      272      269       -3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

mgxd added 3 commits July 17, 2026 16:41
Resample BOLD straight to MNI152NLin6Asym and label the subcortical
grayordinates from the HCP dseg, replacing the per-ROI MNIInfant->MNI6
realignment and removing the abandoned subcortical alignment workflow.
@mgxd
mgxd force-pushed the enh/subcortical-mni6 branch from 2342294 to 28bbaab Compare July 17, 2026 20:51
@mgxd
mgxd marked this pull request as ready for review July 20, 2026 15:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant