We have corrected the things related to "zbuffer" rendering and also used limit option in redhat9 to set the stacksize to unlimited. That did not solve the problem. So I was a little concerned about the model as I remember reading in the archives this also happens when a bad design results in suprathreshold cluster.BTW,I'm using Matlab7.0 SP2 and SPM2

Thanks,

Hima

Benjamin Bender <[log in to unmask]> wrote:

We had a lot of segmentation faults here on different machines (all
running some version of Suse 9.x) during different tasks. We could fix
the problem by disabling z-buffer rendering as suggested for Redhat 8 in
the spm wikibook:
http://en.wikibooks.org/wiki/SPM-Disable_z-buffer_rendering

I cannot answer your other questions though.

best regards
Benjamin Bender

Himachandra Chebrolu wrote:

> Dear Dr. Ashburner and all,
>
> I'm back. I've progressed from the pre-processing stage to the
> statistics stage and I've certainly lerant a lot!Coming to my analysis ...
>
> 1. I've two groups "Male" and "Female" and I would like to do
> something similar to what Good et. al have done with
> VBM(2001NeuroImage:465 Adults). So I'm doing an "ANCOVA" with
> (a) Groups:2
> (b) Nuisance Variables: 1.Age 2.Summed voxels values of gray
>
> The answer we are seeking is " Is there any regional change in gray
> excluding the effects of age and global changes in gray?". After the
> estimation is done and I want to see the regions which are positively
> correlated with global gray calculated and the whole program crashes
> with an error-"Segmentation fault and core dumped".I'm using a
> contrast of [0 0 0 1 0] where '1' corresponds to gray.
> I'm using all default values with the analysis for the grand mean
> scaling etc.
>
> _Questions:_
> 1.Is there something wrong with my design which is causing Matlab to
> crash?
> 2.Also instead of "non-sphericity" correction can I orthogonalise my
> two nuisance variables in matlab and ignore non-sphericity correction?
> What effect will that have on the analysis?
> 3. Why does it asks for a correlated measures e! ven though I only have
> one scan per subject when I choose to do non-sphericity correction?
> 4.Are the replication over groups (or) Subjects?Can you please explain
> the option?
>
> I went through the archives but could not figure these things out in
> the context of my analysis!
>
> Thanks,
> Hima
>
>
> ------------------------------------------------------------------------
> Click here to donate to the Hurricane Katrina relief effort.
>

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com