I used the "cluster" command in FSL to get the information of the significant clusters in the TBSS results. And I found that the clusters were sometimes quite large and not confined to a single white matter tract.
However there are some TBSS studies that show the numbers of the signifcant voxels in each white matter tract (eg. the result of this study are illustrated in this manner: Article White Matter Abnormalities Based on TBSS and Its Correlation...
).So anyone knows how to get this kind of result? The result dosen't look like atlas-based either. Is it possible to use the "fslmaths" or other commands to seperate the clusters from tfce_p_stats map based on white matter tract atlas masks to get this result?
I'm just a beginner to TBSS so any answer is much appreciated!
When you finish TBSS randomise, cluster - can be used to generate information about significant areas cluster(you can also change 0.95 to another values):
cluster -i tbss_tfce_corrp_tstat1.nii.gz -t 0.95 --scalarname="1-p" -o cluster
Atlasquery - can be used to output information on the white matter fibres to which prominent areas belong (two different atlas you can choose one according your require):
atlasquery -a "JHU White-Matter Tractography Atlas" -m cluster.nii.gz
atlasquery -a "JHU ICBM-DTI-81 White-Matter Labels" -m cluster.nii.gz
These are my personal experiences for your reference，please kindly know.
Meiling Gu Thanks for the answer and sorry for the delay of my reply! I've tried the atlasquery command and I've noticed that the results were reported in percentage terms. So perhaps I can multiplied the percentage by the number of voxels in the significant cluster to get the number of voxels in each tract? What do you think?
Liqi Wang There are several auxiliary commands for TBSS result analysis, namely -cluster, -atlasquery,-fslmaths,-fslmeants.
When you use the -atlasquery command it does generate information about the white matter fibre tracts that the significant areas belong to, that is a percentage. Later on you continue to run the -cluster command to split the cluster, and -fslmeants to extract the mean values of the brain areas in the significant areas for other analyses.
If you have any boubt about what i said, you can also contact me by: [email protected].
- Badges
- Science topic
- Similar topics
- Correction