Hi Rosalia, Please see below: On 20 February 2018 at 08:10, Rosalia Dacosta Aguayo <[log in to unmask]> wrote: > Dear FSL's team, > > I have two experimental groups of patients (A and B). Based on previous > findings, I would like to run a Paired One-Group Difference for every group > separately. My question is twofold: > > 1. Is it statistically valid? > It can be valid. > > 2. The model to follow would be this, right? > https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT/UserGuide#Paired > _Two-Group_Difference_.28Two-Sample_Paired_T-Test.29). > It can be if each subject has 2 measurements and you'd like to see if the difference between these two measurements is, on average, different than zero. A paired t-test should not be used to compare groups. All the best, Anderson > > Best regards, > Rosalia > >