Dear Vineeth,
usually, cortical thickness is quite robust to changes in TIV and it's therefore not necessary to control for it. However, there might be data where this additionally removes variance from your data that is related to brain size and there is no clear rule that it should not be considered. Thus, you can use TIV optionally as nuisance parameter it this help to remove variance that is not related to your parameter of interest in your data.
Best,
Christian
On Wed, 16 Mar 2022 10:03:47 +0530, Vineeth Mohan <[log in to unmask]> wrote:
>Dear experts,
>As a beginner in surface analysis, I have a doubt regarding controlling for
>TIV in surface analyses..
>I read in the CAT manual that TIV should not be added as a covariate, but
>couldn't find the reason why. I observed small but significant effects of
>TIV on surface measures of certain ROIs, and also saw some papers where TIV
>was used as a covariate in cortical thickness and gyrification analyses. Is
>it methodologically wrong to include TIV as a covariate in SBM using CAT ..?
>Thank you in advance.
>-Vineeth
>
|