Picking up on an old thread here --
I'm trying to run a melodic analysis on the same group in 2 ways. In the first run I do everything as usual, in the second run I want to apply a correction right after motion correction (using the prefiltered_func_data motion parameters).
From this 2008 post I gather that
- preprocessing in Feat and in Melodic is identical
i.e. is it equivalent to run a full analysis in Melodic, as to do the preprocessings in Feat and then a command-line melodic?
- there is a way to retrieve the script used by Feat
so do the log files produce enough information to reproduce a Feat analysis?
If that is correct then it is possible to do one MELODIC the easy way, and one Feat-preprocessing followed by a command-line melodic.
They should produce the same result (provided I give the random number generaor a push in the right direction).
Afterwards I would be able to use the Feat log files feat2_pre and feat5_reg, insert my correction at the right place in the script, and re-produce the Feat preprocessed images with one line of my own added.
Does that work? Or is there a difference between the preprocessing pipelines of Feat and Melodic? Or is the whole preprocessing pipeline not covered in the logs? (from the feat2_pre log file it seems that work is done in the .feat directory, but the command for entering it is not in the logs)