Statistical analysis with fixel data ModelArray controlling for FDR

1

Hi All,

I had a query regarding statistical analysis of Fixel based datasets.
I am trying to find a way to control false discovery rate (FDR) in statistical analysis instead of the more restrictive control of family wise error (FWE).

I managed to use ModelArray, which is an R based package. However, ModelArray does not incorporate CFE in the current implementation, thus information of fixel-fixel connectivity is not considered in the statistical analysis.

Is there a way to add CFE to the Model Array workflow?
What caveats are there in missing the CFE step from a theoretical perspective?

Any feedback will be much appreciated. Thanks!
Best,
Hasini

2

Follow up query: I have been following this other topic where @rsmith mentioned that other/less conservative corrections such as FDR have not been implemented so far.

The workflow considered for connectivity-based fixel enhancement (CFE) is run via the commands fixelconnectivity and fixelfilter and the smoothed fixel metrics are saved (e.g. fd → fd_smooth). If I used the fd_smooth folder for the ModelArray based statistical analysis, can I essentially do an FDR based statistical analysis with connectivity considered?