I've been toying with FSL for some time now. After having had some
installation issues (non-Redhat-distribution of Linux), I finally got
BUT: it won't give me any results beyond the preprocessing steps.
Some info: I entered a 4d-file containing an event-related,
two-factorial (3x3) experiment (one subject only for the time being)
with 728 volumes, TR=3.02, Highpass=130 as a source. I set up the design
matrix with the 3-columns option (something like "104.3 2 1" and so
forth for each EV; I also tried entering "104.3 0 1" to simulate a stick
function as in SPM2). After that I entered my contrasts (10, but also
tried 2 only)
The rest was pretty much default settings.
In the result.html file there is nothing apart from the usual
preprocessing stuff, the design matrix and the covariance table - no
statistics, no brains, no blobs.
During logging, following lines attracted my attention:
Prewhitening and Computing PEs...
Unthresholding Series... Completed
Saving image... Completed
Thresholding Series... Completed
/home/zap/Work/FSL/fsl-3.1/fsl//bin/contrast_mgr -f design.fts stats
AvwOpen: Error while opening header!: No such file or directory
Rendering using zmin=2.3 zmax=8
/home/zap/Work/FSL/fsl-3.1/fsl//bin/tsplot . -f filtered_func_data
Cannot open volume ./stats/pe1.hdr for reading!
The full log is attached to this mail.
Being aware of those FSL RAM/swap issues, I had increased my
swap-partition to 6 GB while having 512MB of RAM. My system monitor
indicated a full RAM usage and 2.6GB of used swap - leaving some leeway
in my opinion.
Now: what can I do to have FSL run through it all?!? Any help would be
Institute of Medicine
Research Center Juelich