Hi Jasmin, This refers this thread: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=FSL;4e647748.1508 Sorry, I still don't understand what you need or are trying to do. Randomise needs 4D files as inputs, not 3D volumes. Covariates go in the design matrix, for a hypothesis test. What is the hypothesis that you'd like to test? If there isn't one, maybe randomise isn't the correct tool you need? Maybe fslmaths or fsl_glm could do it. All the best, Anderson On 30 September 2015 at 01:50, Jasmin Czarapata <[log in to unmask]> wrote: > Thanks Anderson for that. > > Follow up question: I cannot seem to be able to perform randomize on the > individual volumes (IC) following splitting of each subjects dr stage two > components. > But more importantly each volume does not represent the covariates that is > in the drstage 3 ttest 2. I only have one ev which is covariates scores > (demeaned). In other words the vols do not represent the model which > includes a covariates. The individual volumes are identical to the melodic > IC. Is there a way to get the individual vols (i.e. Vol0002.nii.gz) that > with correlated covariates? > > Thanks. Jasmin >