Hi - one of the internal stats "images" saved relies on a contrastsXcontrasts file - so the requirements to scale up with the number of contrasts as well as timepoints.  If this is a problem, and you are going to rely primarily on ME group-level stats, you could try doing OLS at the first level instead of FILM - though I believe that's not a GUI-controlled option…..

Cheers, Steve.



On 4 Nov 2012, at 15:33, Russ Poldrack wrote:

howdy - I'm having trouble running a first-level analysis in 5.0.1.  This is a fairly large dataset (380 timepoints, 96x96x68 dataset, 72 regressors), and a full feat analysis crashes with an out of memory error when running contrast_mgr on a system with 24 GB of RAM.  The job completes on our large-memory node so it's not a bug, just very large memory footprint, but it seems to me that the job of contrast_mgr should not take that much memory.  any suggestions would be most welcome!
best,
russ



---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------