Hi Michael,

while fslmaths has an option to do "2D" smoothing, this is still only smoothing in a slice of the volume and has the same problems as 3D smoothing (ie, smoothing across tissue types and sulci). The "2D" that FS implements with mris_volsmooth is along the cortical surface, which is unrelated to the volume axes. So, unfortunately, you do still have to do two analyses if you want a volume-based analysis with smoothing.

doug

Michael Scheel wrote:

Hi Estephan,

you don't necessarily have to rerun the full 1st level Analysis in total. What you have to find out is which of your volumes is the one that example_func is created from. Normally this should be the middle timepoint fmri image of your functional run (e.g. if you have 99 volumes then number 45 will be taken as example_func) that's the default if you haven't specified it otherwise. You can use fslroi as a command to extract this volume for the above example e.g.: fslroi input_fmri.nii.gz middle_volume.nii.gz 44 1 (fslroi starts counting from 0 thats why it is 44 instead of 45)

2D and 3D smoothing can also be done with fslmaths see the -s and the -kernel option

Best, Michael On 23-Sep-09, at 6:52 PM, Douglas N Greve wrote:

...

Estephan Moana wrote:

...

So what I need to do is to run the 1st level FEAT as usual for the FSL pipeline, and repeat this analysis for FS with the only difference been to set the "Spatial smoothing FWHM" to zero on the pre-stats tab of the FEAT GUI?

yes

...

Also, can you explain if this smoothing on the FEAT GUI is the 3D you refer to? And how 2D smoothing is accomplished?

Yes, this the 3D smoothing. 2D smoothing on the raw data is more difficult inside of feat. You'd have to use MC and then run mris_volsmooth (FS program) outside of FEAT, then pass FEAT the preprocessed data and tell it not to do MC or smoothing.

...

Thank you.

Estephan

Douglas N Greve wrote:

...

The example_func does not actually get smoothed, so smoothing itself will not affect the registration. 3D smoothing will smooth values from WM and CSF into gray and from across sulci into functionally distinct regions. For this reason we recommend not using 3D smoothing (or doing 2D smoothing on the surface). Unfortunately, this requires two different analyses.

doug

Estephan Moana wrote:

...

Hello all, I took the FS course last June at the OHBM in San Francisco, and am now playing with it to get to know it better. One of the slides on the talk about FSL-FS integration mentioned that the input data (from FEAT) for the reg-feat2anat command should have not been smoothed prior to registering to FS.

Considering that FEAT 1st level analysis has a default of 5mm spatial smoothing (and that's what I used on my data), it seems counterintuitive to need to run a 1st level analysis again so the output files are not spatially smoothed.

So my questions are: 1) having the example_func file spatially smoothed will necessarily hinder a proper registration to the FS anatomical?; 2) if so, should I re-run all my FEAT 1st level analysis without smoothing or is there any other way?

I know this is quite a beginners question, but I could not find the answers on the FS FAQ or the mailing list archives. I appreciate any light on this.

Regards,

Estephan Moana, DDS Graduate Student Oral Biology PhD Program - Neurobiology track School of Dentistry, UNC- Chapel Hill 2140 Old Dental Building, CB# 7455 Chapel Hill, NC 27599 Phone: (919) 966-5680 Fax: (919) 966-5339 Cell: (612) 702-3295 moana@email.unc.edu mailto:moana@email.unc.edu

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422

In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer