Thanks Doug, though I'm still a bit confused:

I'd like to see the smoothed results on the individuals, not just the group - so it seems isxconcat-sess wouldn't be appropriate.

Are you suggesting running mri_surf2surf directly on the contrast sig.nii files? Is smoothing a significance map the equivalent of smoothing the data, then recreating the significance in the sphere space? This didn't seem true to me, which is why I tried to recreate the significance maps as below.

Thanks, -S

Doug Greve wrote:

Hi Scott, you should skip steps 2-6 (that is the old stream). Instead, run isxconcat-sess after #1, then smooth on the surface either in isxconcat-sess or with mri_surf2surf before running mri_glmfit.

doug

Scott Gorlin wrote:

...

Hi,

I'd like some clarification on the recommended procedure for smoothing functional data across the cortical surface in FS-FAST.

I've managed to get it working using the following procedure:

1. analyze my data using selxavg3-sess in the native space, without

smoothing 2) resample the averages into the spherical space using func2sph-sess 3) smooth these averages using sphsmooth-sess (which required several edits to work with .nii files) 4) recreate the contrasts in the smoothed sph space with stxgrinder-sess 5) paint the analysis onto the surface with paint-sess (which also required edits for .niis) 6) view with surf-sess (which I had to edit to read the subject properly)

Surely there's a simpler way to accomplish this? I'm particularly confused because the online documentation says that selxavg3 replaces stxgrinder, but it seems stxgrinder has to be called to make contrasts in non-native space (and uses different code to produce the contrasts than selxavg). Plus, the revision date for stxgrinder is more recent!

Then, it seems odd that isxconcat will reproject the data for a group analysis if it's already sampled on the surface via func2sph.

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

Hi Moe

Could you send us the data ( sig.nii .. directly to my email if it's not big enough ). What steps do you take so that I can mimic the crash here?

Thanks Krish

On Nov 25, 2008, at 3:01 PM, Zeidan, Mohamed A. wrote:

Hi Everyone,

I'm trying to create a label but as I'm highlighting the area of interest in tkmedit, the window closes out and I get the following message:

ERROR: A segfault has occurred. This is not your fault, : but is most likely an unrecoverable error and has : made the program unstable. : : Please send the contents of the file .xdebug_tkmedit : that should be in this directory to analysis-bugs@nmr.mgh.harvard.edu : : Now exiting...

This has happened four times so far. Furthermore, the .xdebug_tkmedit file is nowhere to be found (I checked the directory from which I opened tkmedit). My command line was: tkmedit fsaverage orig.mgz -aux brain.mgz -bc-main .33 7 -overlay $SUBJECTS_DIR/Recall_vol_m_vs_f/Recall/CS+EevsCS-/tal.rfx.m_vs_f/ double/sig.nii -fthresh 1.3 -fmax 3

Anyone else have this problem or have thoughts as to why it's happening?

Thanks so much, Moe

-----Original Message----- From: freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Scott Gorlin Sent: Tue 11/25/2008 2:57 PM To: Doug Greve Cc: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] fs-fast smoothing

Thanks Doug, though I'm still a bit confused:

I'd like to see the smoothed results on the individuals, not just the group - so it seems isxconcat-sess wouldn't be appropriate.

Are you suggesting running mri_surf2surf directly on the contrast sig.nii files? Is smoothing a significance map the equivalent of smoothing the data, then recreating the significance in the sphere space? This didn't seem true to me, which is why I tried to recreate the significance maps as below.

Thanks, -S

Doug Greve wrote:

...

Hi Scott, you should skip steps 2-6 (that is the old stream). Instead, run isxconcat-sess after #1, then smooth on the surface either in isxconcat-sess or with mri_surf2surf before running mri_glmfit.

doug

Scott Gorlin wrote:

...

Hi,

I'd like some clarification on the recommended procedure for smoothing functional data across the cortical surface in FS-FAST.

I've managed to get it working using the following procedure:

1. analyze my data using selxavg3-sess in the native space, without

smoothing 2) resample the averages into the spherical space using func2sph- sess 3) smooth these averages using sphsmooth-sess (which required several edits to work with .nii files) 4) recreate the contrasts in the smoothed sph space with stxgrinder-sess 5) paint the analysis onto the surface with paint-sess (which also required edits for .niis) 6) view with surf-sess (which I had to edit to read the subject properly)

Surely there's a simpler way to accomplish this? I'm particularly confused because the online documentation says that selxavg3 replaces stxgrinder, but it seems stxgrinder has to be called to make contrasts in non-native space (and uses different code to produce the contrasts than selxavg). Plus, the revision date for stxgrinder is more recent!

Then, it seems odd that isxconcat will reproject the data for a group analysis if it's already sampled on the surface via func2sph.

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

---

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

---

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

no, i'm using 4.1 stable. is the development version open to the public? (i don't see anything here: http://surfer.nmr.mgh.harvard.edu/pub/dist/dev-releases/)

Or a patch/some minor edits I can copy into preproc-sess?

Or a *not the easiest* way :)?

thanks again

Doug Greve wrote:

Oh, in that case the easiest thing to do is to smooth the raw data on the surface. If you have a development version of freesurfer, you can do this with preproc-sess using the -surf-fwhm option

doug

Scott Gorlin wrote:

...

Thanks Doug, though I'm still a bit confused:

I'd like to see the smoothed results on the individuals, not just the group - so it seems isxconcat-sess wouldn't be appropriate.

Are you suggesting running mri_surf2surf directly on the contrast sig.nii files? Is smoothing a significance map the equivalent of smoothing the data, then recreating the significance in the sphere space? This didn't seem true to me, which is why I tried to recreate the significance maps as below.

Thanks, -S

Doug Greve wrote:

...

Hi Scott, you should skip steps 2-6 (that is the old stream). Instead, run isxconcat-sess after #1, then smooth on the surface either in isxconcat-sess or with mri_surf2surf before running mri_glmfit.

doug

Scott Gorlin wrote:

...

Hi,

I'd like some clarification on the recommended procedure for smoothing functional data across the cortical surface in FS-FAST.

I've managed to get it working using the following procedure:

1. analyze my data using selxavg3-sess in the native space, without

smoothing 2) resample the averages into the spherical space using func2sph-sess 3) smooth these averages using sphsmooth-sess (which required several edits to work with .nii files) 4) recreate the contrasts in the smoothed sph space with stxgrinder-sess 5) paint the analysis onto the surface with paint-sess (which also required edits for .niis) 6) view with surf-sess (which I had to edit to read the subject properly)

Surely there's a simpler way to accomplish this? I'm particularly confused because the online documentation says that selxavg3 replaces stxgrinder, but it seems stxgrinder has to be called to make contrasts in non-native space (and uses different code to produce the contrasts than selxavg). Plus, the revision date for stxgrinder is more recent!

Then, it seems odd that isxconcat will reproject the data for a group analysis if it's already sampled on the surface via func2sph.

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