Hi Cristiel,

First of all, could you tell us which version of SPM (5, 8, 12?) you're using to perform VBM?

Regarding the steps before actual VBM, you usually do not have to realign images as long as they are all more or less oriented in the classical anterior / posterior commisure (ACPC) orientation (for illustration, see http://airto.ccn.ucla.edu/BMCweb/HowTo/AC-PCpix/acline.gif ). If their orientation is not good enough, VBM will fail and you'll have to reorient the images. SPM (like most neuroimaging programs), requires the images to be aligned as regards to the AC-PC line. The image origin (voxel with coordinates 0 0 0 needs to be the AC; see http://dbm.neuro.uni-jena.de/vbm/vbm5-for-spm5/set-origin/ ).You can do the manual reorientation using the "display" button in SPM and adjusting the image using translations and rotations.

Also, the volume.txt file provides raw volumes (not corrected by brain size). But you can easily compute total intracranial volume (TIV) as the sum of GM, WM and CSF volumes.

Good luck with your preprocessings

Hi Renaud! 

Thank you so much for your response, you actually gave me the answers I was looking for. I'm using the VBM8 on the SPM12.  May I ask, I will use the VBM for comparing the hippocampal volume of two different groups. I have already processed the data necessary in VBM and I'm bit confused about extracting the hippocampus volume. That process seems to variate quite much across different papers.  Do you possibly have any recommendation?

Best regards! 

Hi Christiel,

I don't know the optimal way of extracting hippocampal volume. There are several brain atlases that include the hippocampus - for instance try the Jülich (SPM Anatomy) and Hammers atlases (brain-development.org) which you could use as a mask in the VBM analyses. Or alternatively, use to extract the raw volume. I recommend visiting the SPM forum if you have not done so already. It's on the SPM website, and you can search through their e-mail list archives. There will probably be loads of posts regarding the extraction of hippocampal volume and VBM analysis here.

Hi Lasse

Thank you so much for your answer!

Is there anybody who possibly could help me with these questions. I have been reading a lot and unfortunately I haven't found any article or other source with sufficient description of the procedures.

1. How equivalent is VBM8 respective to the so called "optimised VBM"?

I have been reading several articles and many of the talks about the optimized methodology specially for volume measurements. It seems that the traditional VBM isn't always the best choice for volume measurements.

2. For the VBM preprocessing it seems adequate to create a template in using own data.  That in order to eliminate effects that might occure due to the difference of scanners. How is this performed? Maybe using the VBM8 that isn't needed anymore?

Once again, thank you so much for your help!

Best regards; Cecilio.

Christiel,

This is based on my limited knowledge of VBM, so please take my advice with a pinch of salt. I recommend seeking advice on the SPM and VBM forums as well, as you will get answers from experts there.

1. From my understanding, optimised VBM was developed to overcome some of the limitations of early-days VBM approaches. However, VBM8 now uses DARTEL for normalization, rendering the 'optimised VBM' approach obsolete. So if you use VBM8 (or SPM12, which also uses DARTEL), you'll be fine.

2. With DARTEL, you can generate a study-specific template based on your sample. However, VBM8  uses the IXI template (www.brain-development.org) by default for normalization. You can create your own, through DARTEL, but it would require a large sample size (VBM8 manual says 50-100). See the VBM8 and SPM manuals for more information on creating your own template.

Also, this article might be of use to you: Efficacy of voxel-based morphometry with DARTEL and standard registration as imaging biomarkers in Alzheimer's disease patients and cognitively normal older adults at 3.0 Tesla MR imaging, Mak 2011.

Regards, Lasse