Dear FSL users,
I have a question on reporting TBSS white matter tract changes.
I observed significant voxel change in the TBSS output as shown on fslview but when I extracted the DTI metrics and did statistics on SPSS, no significance was shown.
So, firstly, I ran TBSS on patients compared to controls and got significance on all metrics (FA, MD, AxialD and RadialD) as shown on fslview at p<0.05, corrected.
In order to get numerical values to report, I then created masks of different tracts and extract the values of each DTI metric to compile a data sheet with metrics of each tract of each subject.
For example, to extract FA values from forceps major I used:
fslstats -t all_FA_skeletonised -k Forceps_major_min25.nii.gz -M -S -V -x -X >> FA_ForMaj_min25.txt
As I have compiled a data sheet with metrics of tracts of interest for all subjects, I ran ANOVA on SPSS to look at the significant differences between 3 patient groups and control but recorded different results from TBSS.
I am wondering why this might happen?
One reason might be if using a mask and extract metrics, the value might be an average of the whole tract. For example, if the frontal part of the SLF is more severely affected than the rest of the tract, it will show up on TBSS but in the mask extraction the changes in FA do not differ from controls as much because only a proportion of the tract is affected?
I would very much appreciate if anyone can shed some light on this.
Thanks,
Bonnie
|