Tomas,
> I have a problem running computation part of SnPM 3b with SPM2 and
> Matlab 6.5.1 under W XP. I am able to specify a model (single
> subject, covariate only, 25 subjects, interchangebility block 25,
> image about 5 MB, approximate no of permutations 1000, working
> "planimetrically") but the computation part (snpm_cp command line
> interface) crashes.
Before diagnosing the problem, let me check one thing: If you really
have "25 subjects", then you shouldn't be using the 'SingleSub'
plugin. Rather, you should be using the 'MultiSub, 1 Covariate, 1
scan per subj' plug in.
That said, as you have used the single subject plugin (with
exchangeability block size of 25) you should get the identical
analysis with the single subject and multi-subject correlation
plugins.
> Sometimes it computes a few planes (about 3 minutes for a plane) but
> then there is no progress even with enough memory free (houndreds of
> MB) and processor (Intel Celeron 2GHz) go on working on 100% (and
> system does not freez). Matlab gives no information about what is
> going on. Do you have any explanation for such behavior? Is there
> any way how to check whats wrong?
I don't have Matlab 6.5 anymore on my PC, but I can't reproduce this
problem wtih Matlab 7.1. You seemed to have ruled out the usual
problems (of lack-of-memory).
I don't know why it should behave any differently, but try again with
the multisubject plug in and see if you still have problems. Also, if
you are 'collecting suprathreshold stats', try turning that off.
As for diagnosing the problem futher, before you start the computation
type
dbstop error
and then, when it seems to be hung or taking a very long time, hit
Ctrl-C and you should at least be able to see where in the program you
are.
Let me know how it goes.
-Tom
-- Thomas Nichols -------------------- Department of Biostatistics
http://www.sph.umich.edu/~nichols University of Michigan
[log in to unmask] 1420 Washington Heights
-------------------------------------- Ann Arbor, MI 48109-2029
|