Dear John,

 

Thank you for your reply and this suggestion – much appreciated.

 

Could you please clarify  whether the DARTEL imported for tissues 3 to 6 are needed to create the DARTEL template.

 

@ Christian Gaser: if one is using VBM8, does it make sense to also ask for the DARTEL export affine for CSF and PVE?

 

Best wishes,

 

Stephane

 

-----------------------------------------------------------------------------------------------------

Stephane De Brito, PhD

Birmingham Fellow

School of Psychology

Robert Aitken Building, Room 337a

University of Birmingham, B15 2TT, UK

 

Tel: +44(0)121 414 2641 (internal: 42641)

Web: Dr Stephane De Brito - School of Psychology - University of Birmingham

Current project: http://www.femnat-cd.eu/

 

 

From: John Ashburner [mailto:[log in to unmask]]
Sent: 28 August 2015 12:59
To: Stephane De Brito
Cc: [log in to unmask]
Subject: Re: [SPM] fMRI data: DARTEL exports with VBM8 or SPM12 Segment?

 

As far as I know, nobody has compared the approaches, so I don't think anyone has an answer.  The best strategy is also likely to depend on the properties of your data, so it may be worth experimenting on some previously published data to see what works best.

 

Best regards,

-John

 

 

 

On 26 August 2015 at 17:43, Stephane De Brito <[log in to unmask]> wrote:

Hi SPMers,

I have not had a reply to my post from last week, so I am re-sending it with the hope of getting a few suggestions. Thanks in advance.

Stephane

=========================================================================

Dear SPMers,

I am trying to decide on the ‘best’ pipeline to create DARTEL templates to normalise fMRI data of a large sample (several hundreds) of youths (male and females) aged 9-18 years.

Basically I would like to create DARTEL templates, but I need to decide between using ‘Estimate and Write’ in VBM8 or the ‘Segment’ option in SPM12 to produce the DARTEL export files. I am aware of the advantages that the VBM8 toolbox segmentation offers over the Segment in SPM12, so that’s why I am in a dilemma here. For both options, I would use customised TPMs created using the TOM8 toolbox.

***Using VBM8 I would:
Estimate and write, use my customised TPMs with the low-dimensional: SPM8 default for spatial normalisation. For the writing options I would ask for the following outputs: Dartel export affine for grey and white matter. For a fMRI pipeline, does it make sense to also ask for the Dartel export affine for CSF and PVE? Maybe only the PVE? I assume that these could be used in the ‘Create DARTEL Template’ module.
I am aware that the segmentation in SPM12 has been improved compared to the one in SPM8, so does this mean that using the low-dimensional SPM8 default option rather than Segment in SPM12 would be problematic here even if it just for initial skull-stripping?

***Using ‘Segment’ in SPM12 I would:
Again use my 6 customised TPMs and ask for the DARTEL Imported for grey and white matter. Same as above, does it help to ask for the DARTEL imported for tissues 3 to 6? Also, if understand correctly, here those DARTEL files are rigidly registered segmentations rather than affine-only like in the VBM8 toolbox right? Would this be a reason to go for the VBM8 ‘Estimate and Write’ module rather than the ‘Segment’ in SPM12?

****After the segmentation, I would then create the DARTEL templates and then use the ‘Normalise to MNI Space’ module.

Thanks in advance for your help on this and, of course, I would welcome additional suggestions you may have for the pipeline.

Cheers,

Stephane