 |
 |
Dear Antonio, You probably want to demean your continuous covariates here as your contrast is for the intercept. Kindest Regards Matthew -------------------------------- Dr Matthew Webster FMRIB Centre John Radcliffe Hospital University of Oxford On 11 Jun 2018, at 15:13, Antonio Maffei <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear all, I am writing for an expert advise on the quality of my processing steps. I want to test with Randomise the hypothesis of non-0 activation, while controlling for two covariates. The design matrix looks like this (just the first six subjects for convenience): /NumWaves 3 /NumPoints 30 /Matrix 1 23.90 7 1 28.80 12 1 26.00 9 1 33.60 13 1 29.00 6 1 29.40 11 The first column models the intercept, while the second and the third the effect of the covariates. I've read on this list that if one is interested in testing the slope in a correlation analysis, demeaning the covariate of interest should be necessary, especially in the case in which the intercept is not modeled, but I believe that these concerns do not apply to my case. Am I wrong? The contrast matrix is the following: /ContrastName1 ContrastA /NumWaves 3 /NumContrasts 1 /Matrix 1 0 0 Finally, the call to randomise is the following: randomise -i $input_data -o $output_data -d $design_matrix -t $contrasts -T -1 Is this correct or am I missing something? Thanks for any help and suggestions, Antonio ######################################################################## To unsubscribe from the FSL list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1 ######################################################################## To unsubscribe from the FSL list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1