On 7/24/07, Christian Beckmann <[log in to unmask]> wrote:
> I suspect you did not change the brain/bg % level to 0 in the misc
> tab in feat? If so, the mask will probably be empty, i.e. film never
> gets to see any data.

Hi Christian

Your suspicions were right - I had not done this. Doh.

So I've just redone it with the brain/bg % level set to 0.
However this also fails. I've just been persuing the report.log
file and checking on the outputs more carefully:

example_func is not empty
mean_func is not empty
(new saved) filtered_func_data is empty
mask file is empty

Looks like the 'ip' command takes my 4d input (with data in it)
and returns a new version that is totally empty:

ip filtered_func_data filtered_func_data 0 -m mask | grep THRESHOLD | awk '{
print  }' -

avwstats filtered_func_data -R | awk '{ print  }' -

I'm guessing the problem starts at this point?

Presumably I don't really need to run ip (I guess this is the program that
does spatial and temporal filtering?) so long as I generate my own mask?

(i) create your file (e.g. in matlab)
> (ii) create design.mat and design.con in Feat (set up the stats and
> save the design)
> (iii) run film_gls
> film_gls -rn stat -noest data design.mat 0
> this creates a new directory 'stat' with results in it
> (iv) run contrast_mgr to create the stats images
> contrast_mgr stat design.con
> I've attached an example set of files which work fine for me
> hope this helps
> christian
Many thanks - I'll have a play with these.


Daniel Irlam Ph.D
[log in to unmask]