Dear Cartsten:
Thank you. I did quite some time back. SPM does not save that information, so it is important to
make a note of this. If the model doesn't converge- i.e., show a rate of change of parameters of <
1e-5 then it may not be very good. So if you estimate a bunch of DCM's make sure you save the final
convergence estimate. Also you may need to change the number of iterations allowed for convergence.
This is in spm_nlsi_GN (version 2.19) at line 123. We changed this to 128 though most models
converge in less than 64 iterations. Make sure you ahve all the updates to the DCM code.
regards,
Darren
==============Original message text===============
On Mon, 02 May 2005 8:38:11 am CDT Carsten Giessing wrote:
Dear Darren,
did you receive an answer to your question on the spm-helpline?
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=spm&P=R94243&I=-1(I noticed that the DCM software does not save any information related to
convergence....)
Many thanks,
Carsten
--
Carsten Giessing
Institut für Medizin
Arbeitsgruppe Kognitive Neurologie
Forschungszentrum Jülich
Tel. +(0) 2461/612483
===========End of original message text===========
--------------------------------
Darren R. Gitelman, M.D.
Cognitive Neurology and the Alzheimer's Disease Center
Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
Voice: (312) 908-9023 Fax: (312) 908-8789
--------------------------------
|