sorry for the delay, see below

On 4/14/18 10:00 AM, Boggess, Matthew Jozsef wrote:

Hello,

I just wanted to follow up to see if anyone could help with any of my questions, particularly # 1.

Thank you!

Matt Boggess

---

*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Boggess, Matthew Jozsef MBOGGESS@mgh.harvard.edu *Sent:* Thursday, April 5, 2018 8:02:07 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* [Freesurfer] fmriprep to fsfast interface Hello,

I've been working on taking the outputs of the fmriprep BOLD preprocessing tool (https://fmriprep.readthedocs.io/en/latest/index.html) and coercing them into the fsfast structure so that first levels can be run with mkanalysis-sess and selxavg3b-sess. It's been a bit trickier than anticipated and I was hoping to get some help verifying my current approach (attached to this email).

Question 1:

fmriprep outputs the preprocessed BOLD data in the T1 space (mni305 is not an option). However, the typical spatial normalization workflow uses a registration matrix from the bold to T1 to transform to talairach. Is there a way to do this transformation without having to specify a registration matrix given the functional data is already in the T1 space? My current approach is to use mri_convert with the --apply_transform flag pointing to the talairach.xfm registration matrix in the subject's recon folder, but I am not sure if this is a reasonable workaround.

Do you want to analyze it in the native T1 space or in talairach space? If native, then when you run mkanalysis-sess you can bypass the preprocessing by directly specifying the stem of the functional data you want to process (-funcstem). If tal space, then I think it should work out of the box (it would recalc the registration, but that should  be ok).

Question 2:

fsfast normally calculates the global mean value of the raw data to be used for intensity normalization. However, with fmriprep I only have access to the preprocessed data in T1 space. Would extracting the global mean value for intensity normalization from the preprocessed BOLD in T1 space be valid?

Yes

Question 3:

Is there a way to do first levels in the native subcortical space? If so, what is the filename structure expected (maybe mni305.2mm gets replaced with the subject name as with the surface data)? I've only seen examples transforming to mni305 space, but ideally an option would be to just transfer the preprocessed BOLD in T1 space and run first levels staying in the native T1 space.

See above.

Thanks for your help!

Matt

---

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

Hello,

I just wanted to follow up on this again. Is there any way to transform functional data (already in T1 space) to tal space directly? Something like passing an identity registration matrix to mri_vol2vol or another command that would apply the tal transform without a registration?

If not, I will just try re-registering as you suggested, but I am hoping to minimize the extra processing steps.

Thanks,

Matt

________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Boggess, Matthew Jozsef MBOGGESS@mgh.harvard.edu Sent: Friday, April 20, 2018 8:13:11 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] fmriprep to fsfast interface

Thank you for the response!

I'm interested primarily in tal space, but potentially native as well. For tal space, is there no easy way to just directly transform the data into tal space or provide an identity matrix for the registration?

________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve dgreve@mgh.harvard.edu Sent: Tuesday, April 17, 2018 12:01:46 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] fmriprep to fsfast interface

sorry for the delay, see below

On 4/14/18 10:00 AM, Boggess, Matthew Jozsef wrote:

Hello,

I just wanted to follow up to see if anyone could help with any of my questions, particularly # 1.

Thank you!

Matt Boggess

________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Boggess, Matthew Jozsef MBOGGESS@mgh.harvard.edumailto:MBOGGESS@mgh.harvard.edu Sent: Thursday, April 5, 2018 8:02:07 AM To: freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] fmriprep to fsfast interface

Hello,

I've been working on taking the outputs of the fmriprep BOLD preprocessing tool (https://fmriprep.readthedocs.io/en/latest/index.html) and coercing them into the fsfast structure so that first levels can be run with mkanalysis-sess and selxavg3b-sess. It's been a bit trickier than anticipated and I was hoping to get some help verifying my current approach (attached to this email).

Question 1:

fmriprep outputs the preprocessed BOLD data in the T1 space (mni305 is not an option). However, the typical spatial normalization workflow uses a registration matrix from the bold to T1 to transform to talairach. Is there a way to do this transformation without having to specify a registration matrix given the functional data is already in the T1 space? My current approach is to use mri_convert with the --apply_transform flag pointing to the talairach.xfm registration matrix in the subject's recon folder, but I am not sure if this is a reasonable workaround. Do you want to analyze it in the native T1 space or in talairach space? If native, then when you run mkanalysis-sess you can bypass the preprocessing by directly specifying the stem of the functional data you want to process (-funcstem). If tal space, then I think it should work out of the box (it would recalc the registration, but that should be ok).

Question 2:

fsfast normally calculates the global mean value of the raw data to be used for intensity normalization. However, with fmriprep I only have access to the preprocessed data in T1 space. Would extracting the global mean value for intensity normalization from the preprocessed BOLD in T1 space be valid? Yes

Question 3:

Is there a way to do first levels in the native subcortical space? If so, what is the filename structure expected (maybe mni305.2mm gets replaced with the subject name as with the surface data)? I've only seen examples transforming to mni305 space, but ideally an option would be to just transfer the preprocessed BOLD in T1 space and run first levels staying in the native T1 space. See above.

Thanks for your help!

Matt

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

Yes sorry. I meant without resampling. In that case, I will just re-run as you suggested. Thanks for your help!

________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas N. Greve dgreve@mgh.harvard.edu Sent: Monday, May 7, 2018 1:44:55 PM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] fmriprep to fsfast interface

I'm not sure what you mean to apply the tal without a registration. Do you mean without resampling? We don't have anything that will do that for a 12dof registration.

On 05/04/2018 05:11 PM, Boggess, Matthew Jozsef wrote:

Hello,

I just wanted to follow up on this again. Is there any way to transform functional data (already in T1 space) to tal space directly? Something like passing an identity registration matrix to mri_vol2vol or another command that would apply the tal transform without a registration?

If not, I will just try re-registering as you suggested, but I am hoping to minimize the extra processing steps.

Thanks,

Matt

---

*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Boggess, Matthew Jozsef MBOGGESS@mgh.harvard.edu *Sent:* Friday, April 20, 2018 8:13:11 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] fmriprep to fsfast interface

Thank you for the response!

I'm interested primarily in tal space, but potentially native as well. For tal space, is there no easy way to just directly transform the data into tal space or provide an identity matrix for the registration?

---

*From:* freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Douglas Greve dgreve@mgh.harvard.edu *Sent:* Tuesday, April 17, 2018 12:01:46 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] fmriprep to fsfast interface

sorry for the delay, see below

On 4/14/18 10:00 AM, Boggess, Matthew Jozsef wrote:

...

Hello,

I just wanted to follow up to see if anyone could help with any of my questions, particularly # 1.

Thank you!

Matt Boggess

---

*From:* freesurfer-bounces@nmr.mgh.harvard.edu mailto:freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu mailto:freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Boggess, Matthew Jozsef MBOGGESS@mgh.harvard.edu mailto:MBOGGESS@mgh.harvard.edu *Sent:* Thursday, April 5, 2018 8:02:07 AM *To:* freesurfer@nmr.mgh.harvard.edu mailto:freesurfer@nmr.mgh.harvard.edu *Subject:* [Freesurfer] fmriprep to fsfast interface Hello,

I've been working on taking the outputs of the fmriprep BOLD preprocessing tool (https://fmriprep.readthedocs.io/en/latest/index.html) and coercing them into the fsfast structure so that first levels can be run with mkanalysis-sess and selxavg3b-sess. It's been a bit trickier than anticipated and I was hoping to get some help verifying my current approach (attached to this email).

Question 1:

fmriprep outputs the preprocessed BOLD data in the T1 space (mni305 is not an option). However, the typical spatial normalization workflow uses a registration matrix from the bold to T1 to transform to talairach. Is there a way to do this transformation without having to specify a registration matrix given the functional data is already in the T1 space? My current approach is to use mri_convert with the --apply_transform flag pointing to the talairach.xfm registration matrix in the subject's recon folder, but I am not sure if this is a reasonable workaround.

Do you want to analyze it in the native T1 space or in talairach space? If native, then when you run mkanalysis-sess you can bypass the preprocessing by directly specifying the stem of the functional data you want to process (-funcstem). If tal space, then I think it should work out of the box (it would recalc the registration, but that should be ok).

...

Question 2:

fsfast normally calculates the global mean value of the raw data to be used for intensity normalization. However, with fmriprep I only have access to the preprocessed data in T1 space. Would extracting the global mean value for intensity normalization from the preprocessed BOLD in T1 space be valid?

Yes

...

Question 3:

Is there a way to do first levels in the native subcortical space? If so, what is the filename structure expected (maybe mni305.2mm gets replaced with the subject name as with the surface data)? I've only seen examples transforming to mni305 space, but ideally an option would be to just transfer the preprocessed BOLD in T1 space and run first levels staying in the native T1 space.

See above.

...

Thanks for your help!

Matt

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu mailto: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