Hi,
Yes - files named *tstat1 are related to the first t-contrast,
*tstat2 files are related to the second.
TwoSampT_twoback_tstat1.nii.gz is the unpermuted "raw" tstat image for
the 1st contrast.
Many Regards
Matthew
> I have conducted a non-parametric unpaired t-test in randomise
> comparing two
> groups (15 controls versus 15 patients) using voxel-based
> thresholding.
>
> Contrasts:
> 1.000000e+00 -1.000000e+00
> -1.000000e+00 1.000000e+00
>
> Command line to execute the non-parametric unpaired t-test:
> randomise -i TwoSamp4D.one_back.nii.gz -o TwoSampT_oneback -d
> design.mat -t
> design.con -m mask –x
>
> Output files:
> TwoSampT_twoback_tstat1.nii.gz
> TwoSampT_twoback_tstat2.nii.gz
> TwoSampT_twoback_vox_corrp_tstat1.nii.gz
> TwoSampT_twoback_vox_corrp_tstat2.nii.gz
> TwoSampT_twoback_vox_p_tstat1.nii.gz
> TwoSampT_twoback_vox_p_tstat2.nii.gz
>
> I would like some clarification regarding the output files.
>
> 1. I assume the *1.nii.gz and *2.nii.gz for each output file
> represent the
> controls>patients and patients>controls contrast results respectively?
>
> 2. I understand that I need to consult the *vox_corrp_tstat* files
> in order
> to observe where the two groups differ in the predefined contrasts.
> When,
> for example, I open the *_vox_corrp_tstat1.nii.gz file in fslview, it
> informs me that my maximum intensity value is 0.22738. The randomise
> manual
> says the vox_corrp_tstat outputs are 1-P images, where a value of 1
> is,
> therefore, most significant. Therefore, I am assuming that the
> *_vox_corrp_tstat1.nii.gz file doesn't demonstrate significance as the
> maximum intensity value is simply too low?
>
> 3. What is the TwoSampT_twoback_tstat1.nii.gz file?
>
> Thank you,
>
> Liam.
>
|