Fixel-based Analysis Pipeline

1. Preprocess DWI data : Denoising, Gibbs ringings removal, motion and distortion correction, bias field correction

2. Calculate 3-tissue response functions, representing white matter, grey matter, CSF; then compute group averages of the response functions

3. Upsample DWI data to 1.25 mm voxel size before computing fiber orientation distributions

4. Compute brain masks from upsampled DWI data

5. Estimate Fiber orientation distributions (FOD) with multi-tissue spherical deconvolution

6. Perform joint bias correction and intensity normalization

7. Generate study population-specific FOD template

8. Register and warp all participant FOD maps to population template

9. Compute template mask by warping all DWI masks into template space, then intersect all of them. NOTE : need to make sure all brain masks are not missing any regions, since a single missing region will cause brain mask to be missing that region

10. Compute white matter template fixel mask by using fod2fixel on population FOD map while applying template mask

11. Segment FOD images for each participant to estimate fixels and apparent fiber density (AFD), then reorient to template space

12. Assign participant fixel maps to template fixel mask

13. Compute Fiber cross section metric

14. Can use TractSeg to generate tract segmentations to study specific regions of interest (ROI) and compute average fiber density or cross section in each tract