Hi Jorge,
Thank you very much !! It worked !!
I used the new versions: /lme_mass_F1; lms_mass_fit1
/and/lme_mass_fit_vw1/, which you sent me for the old Matlab.
Additionally, I uploaded in Matlab - /lme_fit_init; lme_FSfit;
lme_Gradient; vec; lme_E1/; /fs_write_fstats; fs_write_Y;/
it ended with:
"/Algorithm did not converge at 0 percent/"
I suppose this means everything went fine. Total time - 90 minutes.
/fstats/ - finished OK
/fs_write_stats/ - finished OK.
I opened the sig.mgh and the results are ok ! :)
Thank you very much for your support Jorge !!
Have a great week !
Sincerely,
Alex
Le 08/12/2012 12:23 PM, jorge luis a écrit :
Hi Alex
Here is a non-parallel version of the vertex-wise mixed effects models.
You should be able to run it in your old Matlab.
Just type
[stats,st] = lme_mass_fit_vw1(X,[1],Y,ni,lhcortex);
and then the inference:
CM.C = [0 0 0 1];
fstats = lme_mass_F1(stats,CM);
Another thing: if you don't see enough signal in your statistical map,
you can smooth the data a bit more eg. FWHM = 15mm. On the other hand,
if you are happy with the amount of signal, then you should keep the
smoothing level at 10mm (so you keep most of the spatial resolution of
the data).
Best
-Jorge
------------------------------------------------------------------------
*De:* Alex Hanganu <al.hang...@yahoo.ca>
*Para:* Jorge Luis <jbernal0...@yahoo.es>
*Enviado:* Viernes 7 de diciembre de 2012 23:59
*Asunto:* Re: Re: [Freesurfer] Longitudinal analysis - contrast
Ok !
The good part - you were right - something changed - the error :)
Now its:
/??? Undefined function or method 'matlabpool' for input arguments ...
Error in lme_mass_fit at 123
Error in ==> lme_mass_fit_vw at 73
/
I will try this again on Monday, on the mac of my colleague, who
has a newer version of Matlab.
Thanks for helping at such a late hour !
Have a nice evening ,
Sincerely,
Alex.
Le 07/12/2012 11:33 PM, jorge luis a écrit :
Hi Alex
I've just discovered what is the problem that you are
experiencing. Your version of Matlab is too old and it is not
very friendly with variables within a parfor loop statement as
newer Matlab versions are. You have either of two choices:
1- Update your Matlab to a 2011 or later version (lme will run
very fast).
2- Go into lme_mass_fit function and change the parfor statement
by a simple for loop (everywhere in the code where you see the
word "parfor" change it by the word "for") and run your analyses
with a single processor:
stats = lme_mass_fit_vw(X,1,Y,ni,lhcortex,[],1);
the analysis is going to be slower but you will have the job done.
You will have to do the same with every parallel function that
you are going to use. For example you are going to use:
CM.C = [0 0 0 1];
fstats = lme_mass_F(stats,CM,1);
for the inference step. Thus, you will have to change every
"parfor" in lme_mass_F by a simple "for".
Hope this will solve your problem.
Best
-Jorge
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
