the results of right hippocampus is worse than the left one

[debin zeng]

Dear ASHS team,  

Thank you for creating this promising tool! 

 

I've completed segmentations for 255 healthy children using the latest fast-ashs and the rated atlas package (N =27, three age groups). The process went smoothly, but when I check the results, I found that many results of the right hippocampus are bad while the left one is better. The result of the right hippocampus often seems  to contain double hippocampus. I don't know why it happened. Can you give me some advice about this issue? 

These are some of the results:

Here is my bash shell:

# align the image to the template given by ASHS
antsRegistrationSyN.sh -d 3 -f "$Atlas_dir"/ref_hm/ref_tse.nii.gz -m "$Brain_dir"/"$Subj_id"/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/T2w_restore_ -n 80 -t a
antsRegistrationSyN.sh -d 3 -f "$Atlas_dir"/template/template_bet.nii.gz -m "$Brain_dir"/"$Subj_id"/MNINonLinear/T1w_restore_brain.nii.gz -o ./"$Subj_id"/hippo-seg/T1w_restore_brain_ -n 80 -t a

$ASHS_ROOT/bin/ashs_main.sh -P -I $Subj_id -a $Atlas_dir -g ./$Subj_id/hippo-seg/T1w_restore_brain_Warped.nii.gz -f ./$Subj_id/hippo-seg/T2w_restore_Warped.nii.gz -w ./"$Subj_id"/hippo-seg/

#transform the label to CHN-PD space
## corr_nogray
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_left_lfseg_corr_nogray.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_left_lfseg_corr_nogray.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_right_lfseg_corr_nogray.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_right_lfseg_corr_nogray.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]

## corr_usegray
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_left_lfseg_corr_usegray.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_left_lfseg_corr_usegray.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_right_lfseg_corr_usegray.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_right_lfseg_corr_usegray.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]

## heur
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_left_lfseg_heur.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_left_lfseg_heur.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]
antsApplyTransforms -d 3 -i ./"$Subj_id"/hippo-seg/final/"$Subj_id"_right_lfseg_heur.nii.gz -r ./$Subj_id/MNINonLinear/T2w_restore.nii.gz -o ./"$Subj_id"/hippo-seg/final/"$Subj_id"_CHN-PD_right_lfseg_heur.nii.gz -n NearestNeighbor -t [./"$Subj_id"/hippo-seg/T2w_restore_0GenericAffine.mat, 1]  

Thank you very much!

Sincerely

Zeng

[debin zeng]

The atlas package is downloaded here: https://osf.io/hrv9n