This one may help you. Callaert D, Ribbens A, Maes F, et al. Assessing age-related gray matter decline with voxel-based morphometry depends significantly on segmentation and normalization procedures[J]. Frontiers in aging neuroscience, 2014, 6: 124.

Thanks Lei, that's of a great help. As far as I know, the longitudinal registration of SPM12 is different from SPM5/SPM8 and is greatly improved. Also, there are currently debates over FSL's non-linear registration (FNIRT). Are there any alternative methods that you're aware of? How is FreeSurfer's longitudinal analysis pipeline?

Hello, Amirhussein Abdolalizadeh,

     Your question refers to an emerging field that clinical geriatrics and neuroscientists are interested in. I think the first question will be focus on the operational conception of "brain atrophy". Till now, there are several indices to measure cortical morphometry, such as gray matter density, gray matter volume, gray matter thickness, surface area etc. 

      FreeSurfer and BrainSuite could analyze the T1 images and demonstrate the results of volume, thickness and surface area. On the other hand, SPM+VBM toolbox could analyze the gray matter density based on different templates.

      From healthy ageing to pathological ageing, brain morphometry changes dynamically. Thus, it is very important to evaluate the longitudinal changes from more than one perspectives.

      I hope these would help you. Good luck!

      Hanna

Hi  Amirhussein

I agree with the answers above. First, brain atrophy can be analysed as  volume reduction, or alternatively, cortical thickness and surface area change. volume based method might be easy to do with VBM, but surface based method might be better in the sense of registration which considers the sulcal and gyral information compared to the volume based registration.

For volume based registration, you might want to use ANTs. It shows good registration accuracy, and people also use it for longitudinal analysis.  

Jidan

Thanks Jidan. For surface-based method do you suggest FreeSurfer?

I've heard of ANTs, do you know or have any papers using ANTs for their longitudinal atrophy analysis? I don't think ANTs itself is able to do statistical analysis, its use has to be combined with another tool.

 I never used ANTs for longitudinal research. Just did a quick search, and this paper might be useful for you in terms of the analytical method: Tessa, Carlo, et al. "Progression of brain atrophy in the early stages of Parkinson's disease: A longitudinal tensor‐based morphometry study in de novo patients without cognitive impairment." Human brain mapping 35.8 (2014): 3932-3944. It is true that if you want to stats you may use other tools after ANTs.

I think Freesurfer is fine with surface-based method since it is widely accepted. 

Thanks Jidan. I searched and checked different approaches in silico and FreeSurfer's longitudinal pipeline is my best option so far.