group template construction with unbalanced group sizes with antsMultivariateTemplateConstruction2

[gjoue]

Hi,

I would like to create a study group template for each type of
anatomical image we have (T1-MPRAGE and T2w-TSE) using
antsMultivariateTemplateConstruction2, to which I will register an atlas
and its labels (with antsRegistration) to be able to automatically
segment my ROIs for each individual and compare group differences.

In this study, there are 3 groups, but the group sizes are not balanced:
60, 80, and 20. We do not expect large anatomical differences in the
structures we are interested in -- at least we do not have any group
with brain damage or huge differences in age. What would be best for
creating the study templates? Including everybody? Or keeping groups
balanced (but that would mean excluding a lot of the data we do have
access to)?

Thanks for any thoughts!
Gina