Dear Anderson,
>It seems the statistic is negative everywhere in one of the contrasts, and
>positive everywhere in the other. This isn't common in a correlation with a
>continuous variable. I'd open the 4D file to make sure it looks fine, and
>check also all the previous steps for possible errors. If the problem
>persists, let us know and we can have a look at the image.
I am looking at the 4D file now and I'm not quite sure if it is ok...unfortunately I'm new to this kind of analysis. However I am trying to repeat the symmetrised skeletonization again and check if something looks different. Can I write you again if nothing changes?
>I'm not sure about what you'd be doing in the last question. Do you mean
>running a correlation in all voxels, then comparing these correlations
>between both sides? It's better to instead do the subtraction of one side
>minus another, then run the correlation, as you are already doing.
Exactly. I agree with you that it is better to run the correlation directly on the difference of the two sides, but I was just looking for an alternative solution due to my problem with this approach.
Thank you very much for your help.
Kindest regards,
Eleonora Mastrorilli
|