 |
 |
Thank you Stephane, yes it was a problem with specifying the model. Eugene, I tried using fsl_regfilt. It looks like most people use melodic_mix to specify the design and then -f to remove the components that look like artifacts. Would it be logical to enter as the design matrix the 6 motion confounds (instead of melodic_mix) and regress them using -f (by specifying the 6 columns we want to remove)? Thanks a lot! Helen ________________________________ From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of Eugene Duff [[log in to unmask]] Sent: Tuesday, April 09, 2013 12:54 PM To: [log in to unmask] Subject: Re: [FSL] FEAT stats tab Hi Helen - Yes, people usually use Melodic, which is pretty good at separating out motion confounds without an explicit model (and may even perform better when you leave the artefacts in full). If you want to regress out the signal beforehand, try running preprocessing only in Feat, then use fsl_regfilt to regress out your confounds. Eugene -- On 9 April 2013 10:47, Stéphane Jacobs <[log in to unmask]<mailto:[log in to unmask]>> wrote: Hi Helen, Yes, it's a version difference. "Add motion parameters to model" is what I was thinking of. However, I hd read your message too quickly, and hadn't realized you were doing resting state analyses. I'm not sure this is the way to go, as you don't have any model to give to FEAT, but I've never done this myself. As far as I know, people usually use Melodic, or other model-free approaches, for such analyses. I'm afraid you'll have to wait for an input from a person more acquainted with RS analyses than me to solve your problem! Good luck, Stephane Le 09.04.13 11:36, Helen Sawaya a écrit : Thank you Stephane, I do not have these tabs on my FEAT. Might it be because I'm using fsl version 4.1.9? I have the option to 'Add motion parameters to model' and 'Add additional confound EVs'. In both cases I need to specify a filename in the full model setup (or else it gives an error). I chose 'Add additional confound EVs' because I might also want to regress out white matter signal as well. Thanks for your help Helen ________________________________ From: FSL - FMRIB's Software Library [[log in to unmask]<mailto:[log in to unmask]>] On Behalf Of Stéphane Jacobs [[log in to unmask]<mailto:[log in to unmask]>] Sent: Tuesday, April 09, 2013 12:15 PM To: [log in to unmask]<mailto:[log in to unmask]> Subject: Re: [FSL] FEAT stats tab Hi Lauren, Feat adds these EVs automatically for you (using the output of Mcflirt) if you select the "Standard Motion Parameters" instead of the "Don't add Motion Parameters" option in the stats tab. Cheers, Stephane Le 09.04.13 10:06, Helen Sawaya a écrit : Dear all, I realise this might be a basic question, yet it is still confusing me. I want to enter the 6 motion parameters as confounds in the stats tab in FEAT. I first entered the matrix in 'Select confound EVs text file' after clicking 'Add additional confound EVs'. But then what do I enter as filename in the full model setup when I choose Custom 1 as basic shape? My analysis is resting state and this is the only matrix I want to input. Your input is much appreciated Thank you Helen