Thank you for your quick response,

I am not doing stats in FEAT (only prestats/registration), but rather doing some precessing outside FSL after filtering, so please bear with me: would that mean going through the "stats" part of feat analysis too, and using "add motion parameters to the model" (I assume I would have to specify a model in the first place though)?

Or are the FEAT residuals stored somewhere I am not aware of, even when I run prestats only?

Thanks again,

Nicola


On 04/26/2012 07:17 PM, Stephen Smith wrote:
[log in to unmask]" type="cite">Hi - yes I think this should be fine - though I suppose that it might work a little better if the .par timeseries were highpass filtered like the data - in which case if you did this via FEAT and took the FEAT residuals, you would get that.
Cheers.


On 26 Apr 2012, at 16:29, Nicola Toschi wrote:

Hi All,

just a quick confirmation: in order to explicitely/only regress out motion parameters from data preprocessed in FEAT, do I just treat the .par file as a design matrix and regress out all 6 columns, i.e. do i run:

fsl_regfilt -i filtered_func_data.nii.gz -d prefiltered_func_data_mcf.par - f "1,2,3,4,5,6" -o outfile

Thanks a lot,

Nicola



---------------------------------------------------------------------------
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
---------------------------------------------------------------------------