Dear list,
referring back to my previous two messages on this topic:
http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0703&L=spm&O=D&F=&S=&P=14113
http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0703&L=spm&O=D&F=&S=&P=23765
I have been able to trace back the "Matrix is singular to working
precision" problem. The problem was due to the erroneous inclusion of
the DCM model of just one single subject which had been estimated with
SPM2 rather than with SPM5, as for the DCM models of all the other
subjects. I apologize for that!
Nevertheless, I am still left with two questions/problems about
spm_dcm_average in SPM5:
1. spm_dcm_average in SPM5 won't work if the DCM models to be averaged
are located in separate directories. All DCM models to be averaged need
to be all in the same directory, otherwise I get an error:
------------------------------------------------------------------------
>> ??? Error using ==> load
Unable to read file
/dati/marcotettamanti/negazione/neg_03/stats_dcm/DCM_modelloP.mat : No
such file or directory.
Error in ==> spm_dcm_average at 52
load(P(model,:),'-mat');
------------------------------------------------------------------------
Is this a bug or a function?
2. Lines 152 to 175 of the spm_dcm_average.m file include a code to
remove all fields with non-averaged subject-specific information from
the DCM_avg.mat. While this seems to be quite logic, this does not allow
one to explore the generated DCM_avg.mat with the DCM 'review' function,
in order to e.g. specify Contrasts of connections.
Is it ok to remove lines 152 to 175 from spm_dcm_average.m just for the
purpose of using the DCM 'review' function?
Thank you a lot for your help!
Greetings,
Marco
--
Marco Tettamanti, Ph.D.
San Raffaele Scientific Institute
Facoltà di Psicologia
Via Olgettina 58
I-20132 Milano, Italy
Tel. ++39-02-26434888
Fax ++39-02-26434892
Email: [log in to unmask]
|