Hello,
I would like to perform seed based correlations on resting state data. I have an idea of what I need to do, but am getting stuck on some of the procedures/steps involved, esp. with getting the data pre-processed correctly so that the filtered func input for fslmeants is correct. Ultimately, I would like to do >>fslmeants -in filtered_func -m ROImask -o timecourseROI, and then correlate the timecourses between different ROIs.
This is what I think I am suppose to do:
a) In Feat: McFlirt motion correction, smoothing 8mm, (no slice time correction), high pass filter cutoff=100s (0.01Hz)
=> is there a way I can select for bandpass filtering (i.e. 0.01Hz - 0.1Hz)?
b) Generation of CSF and WM timecourse (to use as confound regressors). I used FAST to segment CSF and WM, transformed these masks into functional space, where I then used fslmeants to extract their timecourses from the preprocessed data generated in part a).
c) I would also like to generate motion confounds, and can see that I can tick off the 'add motion parameters to model' under the 'Stats' tab in Feat (though, is there a way I can get a text file from McFlirt?). I think I can also input the CSF and WM timecourses generated from b) in 'add additional confound EVs' (and as well as any physio regressor files - cardio, respiratory).
d) The problem is that I'm not sure how to put this all together to generate a filtered func that incorporates all the steps above (motion correction, smoothing, filtering, csf-wm-motion-physio confounds). Feat doesn't run with all the inputs above, unless I enter a model on the 'stats' tab - but I'm not sure whether there is suppose to be a model for resting state data, and if so, how I would model it (i.e. what to fill in the fields for shape, off, on, etc).
I would be very grateful for any advice! Many thanks in advance for your time and assistance.
Cheers,
Joyce
|