Hi
Using Mark Woolrich's suggestion I have now used the read/save avw/nii matlab routines
distributed with FSL to successfully generate artificial 4d data sets with my own timeseries
in them. As a sanity check I thought I would try running FEAT on this new timeseries.
The recipe I followed was:
(a) take timeseries from a single voxel from one subject run, where the voxel was highly
activated in the original (standard whole brain) FEAT analysis. The voxel timeseries was
taken from the preprocessed 4d data (filtered_func_data.nii.gz) from the previous FEAT
analysis.
(b) generate new 4d data set with all voxels set to zero apart from the one timeseries
selected from above
(c) run FEAT on the new 4d data set, with *no* preprocessing (no mc, spatial smoothing etc)
and *no* post-stats.
This fails.If I leave FILM on, it causes FILM to generate an exception and to terminate the
analysis. If I turn FILM off, it generates the following warning:
/usr/local/fsl/bin/smoothest -d 183 -m mask -r stats/res4d > stats/smoothness
WARNING: Extreme smoothness detected in X - possibly biased global estimate.
WARNING: Extreme smoothness detected in Y - possibly biased global estimate.
WARNING: Extreme smoothness detected in Z - possibly biased global estimate.
and although it carries on processing all the files from then on are empty.
So is it possible to do something like what I am trying - to run FEAT on sparse 4d data sets
or (ideally) single voxel timeseries data? If so, how does one do this?
many thanks
Daniel
--
Daniel Irlam Ph.D