Skip to main content

Visual Inspection with FSLeyes

Skull Stripping (Step 2)​

Overlay the extracted brain on the original T1 to check the result:

fsleyes "$nifti_dir/${subj}_struct.nii" \
        "$ants_dir/${subj}_BrainExtractionBrain.nii.gz" \
        -cm blue-lightblue -a 50 &

What good looks like: All cortical gray matter is preserved (including temporal poles and inferior frontal lobe), skull and dura are cleanly removed, and the cerebellum is intact.

Warning signs:

  • Over-stripping — brain tissue removed, especially at the temporal poles or cerebellum
  • Under-stripping — skull or dura still attached
  • Asymmetric stripping — one hemisphere stripped more aggressively than the other

Eddy-Corrected Data (Step 8)​

Scroll through the corrected 4D volumes to check alignment:

fsleyes "$eddy_dir/${subj}_eddy.nii.gz" &

Use the volume slider to step through. Volumes should be well-aligned with no jumping between frames.

Warning signs:

  • Signal dropout — unusually dark slices (the --repol flag should have caught these, but verify)
  • Venetian blind artifact — alternating bright/dark stripes from interleaved acquisition with motion
  • Volume-to-volume jumps — the brain shifts suddenly between adjacent volumes

Optional: FA Maps (Step 9)​

A quick sanity check — not a formal QC step:

fsleyes "$dtifit_dir/${subj}_DTI_FA.nii.gz" -cm hot &

White matter tracts should be bright, CSF should be dark, and there should be no streaks or rings. If the FA map looks wrong, the problem almost always originated at an earlier stage.