Print

Print
Hi - NumContrasts  should simply reflect the number of contrasts (rows in the contrast matrix), which doesn't directly depend on the number of EVs in the main design matrix - so if I understand you correctly you don't need to change this.

Cheers.



On 29 Jan 2014, at 08:34, Elisa Veronese <[log in to unmask]> wrote:

Dear FSL users,
I have 2 groups (patients and HCs) and I'd like to run a simple 2 sample t-test with more than one (let's say 3, age, sex and QI) nuisance variable.
After running all the tbss scripts, I've created the design .mat and .con files with the design_ttest2 command.

With a text editor, I've added THREE mean-centred columns in the .mat, under Matrix, one for each nuisance variable.
I've also changed NumWaves to 5, and added three values to PPheights.

In the .con file, I've added THREE 0 to the end of each contrast, changed NumWaves to 5, and added three values to  PPheights.

Now my question is: if I want to look for differences between patients and HCs without being misled by age (or sex or QI),  should I also change the NumContrasts field in the .con file?

Thanks.

Regards,
Elisa


---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------

Stop the cultural destruction of Tibet