Open slamballais opened 3 years ago
Hi,
Is this something that is now do-able with more recent updates? Thanks!
Hello,
This has not been implemented and it is not a top priority as it stands, sorry. I gave it a few tries in 2021 and 2022, but the main issue is that any attempts to implement it in the current code led to unrealistically slow results. I'm leaving the topic because I still would like to see it implemented, but it would require a good use case to justify the effort it takes to implement.
Right now, the brain can only be specified as the outcome. Ideally, we would also be able to model the brain as a determinant.
The existing problems entail: 1) Efficient iterative computation of the X side of the regression formula 2) Proper adjustment for multiple testing
The second point probably(?) requires us to perform permutation testing, which in turn would require us to speed up QDECR considerably. Still, this should be feasible to do.