Hi all
I'm creating a script to automatically analyze an event-related experiment.
before applying FEAT stats/post stats, I would like to perform a melodic denoising using the timeseries model and contrast (con/mat files).
Since paradigm stimuli onset differs among subjects, I don't have ONE con/mat pair, but each subject must have their own files.
To let you understand my method I describe my processing steps for an automatic full analysis with feat:
1) I create the model with the GLM GUI. specifying EV, contrasts and events onset files (for a certain subject) and i save it (model.fsf).
2) i open feat gui, i import the previous file (model.fsf) I define all other options (pre-stats, registr, post-stas) for a certain subject and I save the settings (full_analysis.fsf).
3) I manually edit the full_analysis.fsf, replacing "/home/data/..../path to (e.g.) fmri data" with "fMRI_DATA_PATH" string
4) in the scripts, first I substitute fMRI_DATA_PATH with subject path to (e.g.) own fmri data. and so on for all the other subject dependent parameters. i create subject_AAA_full_analysis.fsf
5) i run: feat subject_AAA_full_analysis.fsf
the same process works for melodic analysis without model, I create a full melodic.fsf file with melodic gui and i substitute subject-dependent information by script. then i run the same command: feat subject_AAA_full_melodic_analysis.fsf
so.....for melodic+model analysis, i need the command that read the model.fsf (after having filled in the correct subject events onset files) and return the corresponding mat/con files.
then i fill in these tho file paths in the subject_AAA_full_melodic_analysis.fsf file and all should work fine.
thanks in advance
Alberto
|