Hi Anderson,
Thank you very much to the quick reply!!
I have DTI and PET data for 16 patients and 12 healthy controls.
I ran TBSS analysis to study the differences between the groups in FA. I want to investigate the PET images using TBSS pipeline for non-FA maps.
The PET images are smoothed (Gaussian-5 mm), and the DTI data is not smoothed ( it is an output of FSL/FDT processing pipeline).
I aim to study the changes in FA values and the changes in PET signal between the two groups using TBSS.
I am not sure what is the optimal solution for the smoothed images. But I am thinking:
1. Either to smooth the DTI maps, similarly to the PET images.
2. Or to use the non-smoothed PET images.
Personally I don't prefer smoothing the DTI data as no enough studies have been done on eigen smoothing http://www.rspq.org/pubs/ep.pdf
May be using non-smoothed PET images is a solution!! but I am not sure
I highly appreciate any advice or input on this
Thank you in advance
Jon
Hi Jon,
It's possible to apply smoothing but generally one prefers less smoothing. Why do you think it would be useful?
All the best,
Anderson
On 30 July 2016 at 12:43, Jon Anderson <[log in to unmask]> wrote:
Dear experts,
Can guassian smoothing be used on FA maps?
I have diffusion maps ( FA, MD, AD, and RD) derived from the FSL pre processing pipeline using the tool ( FDT).
I am thinking of applying the following command on FA, MD, AD and RD
fslmaths <DTI_map.nii.gz > -kernel gauss 2.55 -fmean <DTI_map_sm6mm> -odt float
Can smoothing be used on DTI maps then including these smoothed maps in further analyses (e.g. TBSS)
Thank you
Jon
|