 |
 |
Hi Andreas, Hmm.. The equivalent of demeaning here is to remove the mean RT for each subject from the RT regressor. This will only affect contrasts looking at subject means, not at RT, and I think these are not of interest here, so this is probably not really necessary. Cheers Eugene -- Centre for Functional MRI of the Brain (FMRIB) | University of Oxford John Radcliffe Hospital | Headington OX3 9DU | Oxford | UK Ph: +44 (0) 1865 222 523 | Mob: +44 (0) 7946 362 059 | Fax: +44 (0) 1865 222 717 -- On 28 September 2010 13:02, Andreas Bartsch < [log in to unmask]> wrote: > Hi Eugene, > No demeaning? > Hope all is well. > Cheers- > Andreas > ________________________________________ > Von: FSL - FMRIB's Software Library [[log in to unmask]] im Auftrag von > Eugene Duff [[log in to unmask]] > Gesendet: Dienstag, 28. September 2010 10:44 > An: [log in to unmask] > Betreff: Re: [FSL] paired t-test with behavioural correlates > > Hi Patrick - > > Yes, that's reasonable way to model a covariate in this design. > > Regards, > > Eugene > > -- > > > On 28 September 2010 00:34, Patrick Thompson <[log in to unmask]<mailto: > [log in to unmask]>> wrote: > Hi All, > > I have a question on statistics. I have scanned (perfusion scan) multiple > subjects with each subject scanned twice (before treatment-after treatment). > I am interested in looking for a difference between two scan session and any > correlation with behavioural changes. I know, I need to conduct a paired > t-test. However, I am just wondering on how to include other behavioural > measures including average reaction time etc in the paired t-test model. > > For example, if I have two subjects scanned under two conditions A and B, > is this the right design matrix with RT as a covariate of interest. > > > A>B S1 S2 RT > 1 1 0 0.3 > -1 1 0 0.1 > 1 0 1 0.5 > -1 0 1 0.2 > > contrast A>B 1 0 0 0 > contrast B>A -1 0 0 0 > RT>0 0 0 0 1 > RT<0 0 0 0 -1 > > Can I include other covariates (that are not correlated) in this model ? > > > Regards, > Pat > >