Hi
If the melodic dim is too high then you might run out of memory, e.g. if you have a dim of 50 you're already way into the 13000 images when concatenating…
If all else fails you could run group ICA on a random subset and then run dual regression against the full set of subjects.
hth
Christian
On 16 Aug 2011, at 23:12, Saemann Philipp wrote:
> Dear FSLers,
>
> I am using Melodic (latest version) for time-concat. group ICA getting
> an error message regarding memory allocation (on a 64 bit Linux machine),
> while there should be sufficient space on the disc, and the swap space
> has been extended to 16 GB.
>
> The job is interrupted before the PCA step starts.
>
> The resolution of the data is 3x3x3 mm3 with 150 images per subject,
> and N=245 subjects.
>
> Do any defaults have to be changed in FSL scripts? Otherwise I'd have
> to check more computer specific problems,
>
> Thanks for any advice on this!
> Philipp
> Max Planck Institute of Psychiatry
> NMR Research Group
> Kraepelinstr. 2-10
> 80804 Munich
> Mail: [log in to unmask]
> Phone: 0049-89-30622-413
|