Hi!
I am doing an analyses on two groups of subjects with data that was acquired with different scanning parameters:
1. TR 2.3, 37 slices, 3x3x3
2. TR 2.3, 34 slices, 4x4x4
What would be the best common voxel size for the analyses?
To what point should I preprocess the data separately?
Thank you for any hepl in advance,
Emilia Cheladze
Forums
From the data, I would choose
From the data, I would choose 4x4x4 for comparison. Although 3x3x3 is the most common resolution used.
Best,
Chao-Gan
Thank you very much for your
Thank you very much for your answer!
If you don't mind I have few follow up questions, though.
1. What exactly happens in DPRSFA when I input the voxel size during Normalization step? Is the image actually resliced (if the input voxel size is different that it actually is) ? If yes, than could you please write what is the interpolation method?
2. If I were to reslice image in utilities>image reslicer when is the best moment to do it? Before/after preprocessing, first level?
3. What the choice of the interpolation method depends on? What would be the one you recommend?
Thank you very much for your
Thank you very much for your answer!
If you don't mind I have few follow up questions, though.
1. What exactly happens in DPRSFA when I input the voxel size during Normalization step? Is the image actually resliced (if the input voxel size is different that it actually is) ? If yes, than could you please write what is the interpolation method?
2. If I were to reslice image in utilities>image reslicer when is the best moment to do it? Before/after preprocessing, first level?
3. What the choice of the interpolation method depends on? What would be the one you recommend?
1. DPARSFA call's SPM's
1. DPARSFA call's SPM's normalize function. You can find the details at SPM's manual. The interpolation is 4th Degree B-Spline (the default one).
2. You don't need to use Image Reslicer if you set the voxel size in Normalization.
3. I will choose the default one.
Thank you very much for your
Yes.
Yes.
However, you have to resample the 4x4x4 data in the normalization step, though keep the same voxel size.