I previously ran some FEAT models on data from older adults (65+) but noticed
that for those with very large ventricles and atrophy the registration was not
very good. So I got a reply on here to try Nonlinear warp resolution on the
registration tab and it works very well (Thank you, Stephen).
However, it takes terribly long to do even one run for first level. We have like
4-5 different models (and 50-60 subjects) that we now want to re-do. I've
completed one model for all subjects and am wondering if there is a way to
bypass the registration for other models by using something from this
completed model (kind of similar to using preprocessed data from one model,
like filtered_func_data.nii.gz, as input in subsequent models to bypass
prestats).
Because each model regroups the data differently I wasn't sure if this was
possible but any advice would be appreciated.
Thanks,
Kevin
|