Hi
Yes, if you use basis functions current FEAT won't like it - we have
changed this for the next release version so that you can directly
specify additional confounds (won't even need mccutup anymore). In the
meantime the only simple option I can think of is to switch to simple
convolution rather than basis sets. If, however, you think you do
need basis sets then one trick is to specify the design without
confounds first (using flobs), save this and then extract the flobs
regressors from design.mat.
You can then save these as text files and re-create the design with
these using a 1 column shape and no convolution, now including both
the EVs generated from flobs and the confound EVs from mccutup... bit
of a pain but scriptable ;)
hth
Christian
On 29 Jul 2008, at 13:25, Heather L. Urry wrote:
>
> Hi Christian,
>
> Thanks for the suggestion-- I didn't know about the mccutup script,
> and it'll be quite useful I'm sure. Unfortunately, this doesn't
> work... when I add the motion parameters (e.g., mc1.par), I get an
> error message indicating "In order to set up contrasts in "Original
> EVs" mode whilst using basis functions or sinusoidal harmonics, all
> the original EVs must be the same basic shape and generate the same
> number of real EVs. Please change the EV setup." I would gather that
> this is occurring because all the other EVs use the 3-column format
> file and/or because all the other EVs are convolved with the default
> flobs whereas the motion parameters are not (?). In this case, shall
> I go ahead and run the Pre-Stats to get FEAT to create the motion
> parameters (with no other Pre-Stats options selected) and then run
> FEAT again to Stats/Post-Stats?
>
> Next time I'll run MELODIC within FEAT... *sigh* Oh well. :)
>
> Cheers,
>
> Heather
>
>
> On Mon, Jul 28, 2008 at 5:54 PM, Christian F. Beckmann <[log in to unmask]
> > wrote:
> Hi
>
> The easiest thing to do is to run mccutup on the ica directory. In
> FEAT, you'd then run with 'Stats/poststats only' and add 6 more EVs
> in the design GUI, choose the 'custom 1 column' EV type for each one
> of them and specify the 3 rotation and 3 translation time courses.
> hth
> Christian
>
>
>
> On 28 Jul 2008, at 22:28, Heather L. Urry wrote:
>
> Hello all:
>
> We're using MELODIC to denoise at the single subject level prior to
> running FEAT. As such, we did the Pre-Stats steps (motion
> correction, high pass filtering, etc.) as part of MELODIC (i.e., not
> in FEAT), identified the components to remove, and then used the
> fsl_regfilt command to remove them from filtered_func_data.nii.gz,
> saving denoised_data.nii.gz. I was next going to feed this
> denoised_data file into FEAT for the Stats and Post-Stats steps (Pre-
> Stats turned off), and I'd like to include motion parameters in the
> GLM. However, with Pre-Stats turned off, the "add motion parameters
> to model" option in the Stats tab is unavailable. I presume this is
> because Feat will assume that there are no motion parameters to
> model, but of course there are-- they're in the MELODIC directory. I
> was thinking I could tell FEAT to run a Full Analysis, but deselect
> all Pre-Stats options since this seems to make available the "add
> motion parameters" in the Stats tab, but how can I tell it where to
> find the motion parameters? Or, do we need to run the Pre-Stats
> (motion correction only) on the original data so that the motion
> parameters are generated anew, and then run Stats and Post-Stats?
>
> Thanks in advance for your input!
>
> Cheers,
>
> Heather
>
>
> --
> Heather L. Urry
> Department of Psychology
> Tufts University
> 490 Boston Avenue
> Medford, MA 02155
>
> email: [log in to unmask]
> phone: 617-627-3733
> fax: 617-627-3181
>
>
>
> --
> Heather L. Urry
> Department of Psychology
> Tufts University
> 490 Boston Avenue
> Medford, MA 02155
>
> email: [log in to unmask]
> phone: 617-627-3733
> fax: 617-627-3181
|