Hi,
Did you start Feat with the GUI after changing the alternative mask
file, or did you try to run it from the command line ('feat design.fsf').
I use the alternative mask option as well and did not notice obvious
problems. In the log file you should find something like :
/usr/local/fsl/bin/fslmaths prefiltered_func_data -mas
/DATA/Data_procs/scan_04/scan_04_pp_brainmask.nii.gz
prefiltered_func_data_altmasked
/usr/local/fsl/bin/fslstats prefiltered_func_data_altmasked -p 2 -p 98
0.000000 1763.737915
/usr/local/fsl/bin/fslmaths prefiltered_func_data_altmasked -thr
176.3737915 -Tmin -bin mask -odt char
However, I allways run feat from the command line, having the bet option
deactivated.
I hope this helps a bit,
wolf
Marios wrote:
> Hi all,
>
> I've been having some trouble getting the alternative_mask non-gui option at
> the bottom of the design file to actually take effect. I tried two things: I
> ran a full first level analysis by specifying an alternative mask in the fsf
> file, while 1) leaving the BET option turned on and 2) de-activating the BET
> option in the Pre-Stats (just in case the two things somehow conflict).
>
> Unfortunately, the alternative mask was never used. As was mention in a
> previous post, I would expect the resulting mask file (i.e. mask.nii.gz in
> the .feat directory) to be identical to the one requested in the design
> file. Instead, from what I can see from the log file, fSL ignores this
> option and uses either betting or regular thresholding to generate the mask
> depending on whether BET is on or off.
>
> Is it possible to make this option work? Perhaps hard coding it in the
> feat.tcl file might work (though I haven't tried this yet) but it would be
> an inconvenience given that for every subject I would need to use a
> different alternative mask. Any help would be greatly appreciated.
>
> Thanks
> Marios
>
|