I created a T1-weighted and FA Template from my subjects using buildtemplateparallel.sh in ANTS. Now, I want to create a template FOD that matches the template initially generated. I have tried using population_template before. However, my worry is that since the population template creates its own template, it might not match the initial templates I have. Is there a way to work around this? Is it advisable to register the output of the population template to the FA Template?
Take the nonlinear warps generated through ANTs template construction, apply those transforms to the FOD images using MRtrix3, and compute the mean of these images to produce the template FOD image.
The final FOD template image produced by population_template is just the result of transforming each input FOD image according to the final estimated transform and computing the mean. So if you have existing transforms computed by some other package, you can just manually perform those final steps yourself without ever invoking population_template.
The advantage of this approach is that the resulting image will intrinsically align with your existing multi-modal template. A potential disadvantage is that the FOD information will not have been used as part of the template construction and registration process, so if use of FA led to misalignment of structures within WM, this will be reflected in the FOD template.
Generate an FOD template independently using population_template, and then perform registration between that FOD template and the existing multi-modal template to be capable of transforming one to the other.
Would require some careful thought regarding exactly how to do registration between FOD data and non-FOD data. Not to say it can’t be done, just there isn’t a single unambiguous answer for how to do it.