Do you have data from a few subjects scanned at each site? For example, 5 subjects each scanned at each site? If yes, then you could compare the FA between the scans to validate whether the results are the same. If the FA values are the same then it would be ok to combine the data from different sites. However, different b-values will generate different diffusion encoding sensitivity, and each b-value sequence will probably have a different TE. But, there are quite a few papers showing how to use multisite DTI data, such as:

https://www.ncbi.nlm.nih.gov/pubmed/21316471

https://www.ncbi.nlm.nih.gov/pubmed/23629049

https://www.ncbi.nlm.nih.gov/pubmed/24657781

https://www.ncbi.nlm.nih.gov/pubmed/25026156

https://www.ncbi.nlm.nih.gov/pubmed/26872408

https://www.ncbi.nlm.nih.gov/pubmed/27330240

https://www.ncbi.nlm.nih.gov/pubmed/27350723

https://www.ncbi.nlm.nih.gov/pubmed/28190639

Jerome 

Hi,

could/should add different sites as (dummy-coded) covariates in any regression analysis you perform and of course include/discuss as limitations.

You could further assess whole brain similarity of FA values e.g. compared to an FA template (relative values across brain regions) or across sites (absolute voxel-wise values).

Hope this helps & good luck!

Hi, 

I also performed a mulicenter study using DTI for several different centers. I think that it is fundamental to obtain data from each group investigated (e.g. controls and patients, etc.) from each center, and to correct the analyses for center. Clearly the optimum could be to acquire in a short period of time some controls in each site, but I think that this approach is often very complicated.