On 22 Jan 2013, at 22:54, Anders Eklund wrote:
Dear FSL experts,
through the FSL logs I have managed to write a bash script that runs the first level analysis. However, for the higher level analysis it seems like some commands are missing in the logs. I understand that the results from the first level analysis are transformed into MNI space and concatenated into one cope and varcope file for each contrast, but I can't see the actual call to flame. In another log file I found the following command,
flameo --cope=filtered_func_data --vc=var_filtered_func_data --dvc=tdof_filtered_func_data --mask=mask --ld=stats --dm=design.mat --cs=design.grp --tc=design.con --runmode=fe
but I can't find how filtered_func_data and var_filtered_func_data are calculated, are these simply equal to the cope and varcope files?
Yes.
However, I strongly recommend not trying to replicate lower-level or higher-level FEAT via a bash script as there is a very good chance you will not end up doing exactly the same as what FEAT carries out. As you rightly say, some things happen within the TCL and are not logged, for example. The log is not intended to allow a simple reconstruction of the whole analysis from a new bash script.
Given that it is straightforward to setup design.fsf files from the GUI and amend them with scripts, and then call the command-line version of feat, I would strongly recommend doing that instead, if you are wanting to (e.g.) do lots of runs from the command line.
Cheers.
It would be great if anyone could tell me which source file to look into.
Cheers,
Anders
