Index out of bound during Tractometry

I am getting an error when running tractometry that reads: UserWarning: Thedipy.stats module is still under heavy development and functionality, as well as the API is likely to change in future versions of the software warnings.warn(w_string) 0%| | 0/50 [00:00<?, ?it/s] Traceback (most recent call last): File "/usr/local/bin/Tractometry", line 138, in <module> main() File "/usr/local/bin/Tractometry", line 109, in main predicted_peaks=predicted_peaks, affine=scalar_image.affine) File "/usr/local/lib/python3.5/dist-packages/tractseg/libs/tractometry.py", line 64, in evaluate_along_streamlines streamlines = _orient_to_same_start_region(streamlines, beginnings) File "/usr/local/lib/python3.5/dist-packages/tractseg/libs/tractometry.py", line 42, in _orient_to_same_start_region if beginnings[int(startpoint[0]), int(startpoint[1]), int(startpoint[2])] == 0: IndexError: index 178 is out of bounds for axis 1 with size 174

Steps taken to replicate on TractSeg 2.1.1: 0) Preprocessed dwi (Diffusion.nii.gz, single shell acquisition, has undergone denoising, eddy, and N4 bias correction), along with the b0 mask (nodif_brain_mask.nii.gz) and FA map (FA.nii.gz) are copied to the working directory.

1. Create peaks from preprocessed data (not rigidly aligned to MNI, but in the same orientation) TractSeg -i Diffusion.nii.gz -o tractseg_output --raw_diffusion_input --brain_mask nodif_brain_mask.nii.gz --output_type tract_segmentation

2. Remove the recently created bundle segmentations because I have to flip the peaks over X axis (discovered this primarily through trial-and-error). Is there a wrapper function in TractSeg just to produce the MrTrix peaks so I can avoid creating the bundles twice? rm -r tractseg_output/bundle_segmentations

3. Flip peaks and recreate bundle segmentations flip_peaks -i tractseg_output/peaks.nii.gz -o tractseg_output/peaks_flip_x.nii.gz -a x TractSeg -i tractseg_output/peaks_flip_x.nii.gz -o tractseg_output --output_type tract_segmentation The peaks and bundle segmentations look correct after this step (can provide images if needed)

4. Create Ending Segmentation, TOM, and TOM trackings TractSeg -i tractseg_output/peaks_flip_x.nii.gz -o tractseg_output --output_type endings_segmentation TractSeg -i tractseg_output/peaks_flip_x.nii.gz -o tractseg_output --output_type TOM Tracking -i tractseg_output/peaks_flip_x.nii.gz -o tractseg_output --tracking_format trk (should I specify --nr_fibers here?)

There are no errors up until this point, and the .trk fibers look great!

1. Attempt tractometry Tractometry -i TOM_trackings/ -o FA.csv -e endings_segmentations/ -s FA.nii.gz

This is where the error occurs. If it’s of any help, I am using the Brainlife singularity container . Please let me know if anymore information would be helpful.

Thanks, Steven

[smeisler]

Looking through the documentation of Tractometry, I did not know that there was a --tracking_format argument for that as well. I changed it to trk to match the format of Tracking, and now it seems to work.

[smeisler]

The following outputs from the fslhd command show the sform and qform affine matrices. While the offset did not seem to be in those matrices, there was an image_offset output which was constantly 352.

FA.nii.gz

OFFSET: 352

peaks_flip_x.nii.gz

OFFSET: 352

bundle_segmentations/CST_right.nii.gz

OFFSET: 352

endings_segmentations/CST_right_b.nii.gz

OFFSET: 352

TOM/CST_right.nii.gz

OFFSET: 352