 |
 |
I'm sorry to not have the message I was replying to attached(see below now), so here is a repost. I have done similar analysis as in the post below. However, when looking at the outputs of contrast 1 and 2 (or 3 and 4), they are not the same. How can this be? If a regression is multiplied by -1, I should get similar P-values right? Thanks, Sjoerd On 3 Jun 2011, at 04:29, R Duke wrote: > Dear FSL Experts, > > > > I have a question regarding my design matrix. I have two groups (G1, G2), one covariate of interest (CI), and three covariates of no interest (CNI). I am interested in within group correlations with my single covariate of interest and was wondering if the following contrasts were correct for the design matrix I have listed below. > > > > Contrast1: 0 0 1 0 0 0 0 0 0 0 (give me voxels that show a positive correlation with my covariate of interest in group 1) > > Contrast2: 0 0 -1 0 0 0 0 0 0 0 (give me voxels that show a negative correlation with my covariate of interest in group 1) > > Contrast3: 0 0 0 1 0 0 0 0 0 0 (give me voxels that show a positive correlation with my covariate of interest in group 2) > > Contrast4: 0 0 0 -1 0 0 0 0 0 0 (give me voxels that show a negative correlation with my covariate of interest in group 2) > On Oct 21, 2011, at 1:27 PM, Sjoerd Bruijn wrote: > Hi FSL experts, > I have done similar analysis as in the post above. However, when looking at the outputs of the -1 and 1 contrasts, they are not the same. How can this be? If a regression is multiplied by -1, I should get similar P-values right? > Thanks, > Sjoerd Sjoerd M. Bruijn, PhD Motor Control Laboratory Research Center for Movement Control and Neuroplasticity Dept. of Biomedical Kinesiology K.U.Leuven Tervuursevest 101 3001 Leuven [log in to unmask] http://www.move.vu.nl/members/sjoerd-bruijn/