This means that the design matrix is ill-conditioned. This can happen if your using an FIR model (ie, not using -gammafit) and the stimuli are presented with some fixed timing with respect to each other with the elapsed time between them being less than the FIR time window. For example, if you presented a stimulus every 16 sec but had a 20 sec time window. Or if one stimulus is always followed by another with a fixed delay. I suspect the latter is the case as it's hard to avoid with a memory task. Unfortunately, you're pretty much forced to assume a shape to the hemodynamic response (ie, use gammafit).
doug
Olivier Piguet wrote:
Good morning,
I am getting the following error message whilst trying to run the selxavg-sess command:
ERROR: paradigm is ill-conditioned (5.37345e+18). Check your paradigm file for file for periodicities or for some event types that all ways follow other event types (or itself).
quiting matlab
ERROR: fast_selxavg() failed
ERROR (/autofs/space/engram_010/users/rk_fmri_test/YNS15_060404): selxavg failed
What is the meaning of this message? I have checked the mail archives but couldn't find any information.
I am analysing data for a retrieval memory task and the para file (number of conditions and response types) is based on the subject responses and therefore some response types may or may not be repeating themselves.
Thanks
Olivier
--
Olivier Piguet, Ph.D. Massachusetts Institute of Technology Department of Brain and Cognitive Sciences NE 20-392 77 Massachusetts Avenue Cambridge MA 02139 USA
For overnight or express deliveries, send to: 3 Cambridge Center Suite 392 Cambridge, MA 02142 USA
Voice: (+ 1) 617-253-8976 Fax: (+ 1) 617-253-1504 E-mail: olivier@mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer