Dear Maaly,
yes, you're better off changing SPM defaults otherwise you will come
across the same problem everywhere in the code where the SPM.mat file is
saved.
Sorry if your FFX model is very slow to specify/estimate and uses a lot
of memory - SPM is far from being optimised for these models. Why are
you doing a fixed effect analysis in the first instance?
Best regards,
Guillaume.
On 04/07/16 12:43, Maaly Nassar wrote:
> Dear Guillaume and Rouhollah
>
> Rouhallah's approach worked for the model estimation but not for the contrasts (see attached images), and accordingly, the model estimation was deleted (replaced by empty spm.mat after the contrasts error :( ). So I run my batch script for FFX as sections and stop at model estimation then I modified spm_contrast.m line 313 and since 10:00 today till now it did only 3 contrast images, so why it is too slow? I also tried to search for the spm_defaults file and changed line 49 as you hinted and run the script and FFX specification took more than 30 min to save configurations, so I had to stop the script. I'm still waiting now for the "contrast manager" batch. I'm just wondering about how long will it take if I had 20 subjects instead?
>
> But despite all this, thank you so much for both of you. At least it is working
>
> maaly
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|