[pymvpa] Pkg-ExpPsy-PyMVPA Digest, Vol 16, Issue 8

[Kemal]

>> 1. When I try to follow the tutorial
>> (http://www.pymvpa.org/examples.html), I get an error right at the
>> beginning, when reading in the sample dataset (bold.nii.gz). The error
> which particular example runs out of memory?
>
> we run daily tests on a box with 3GB -- never had a problem with
> examples -- example dataset shipped within pymvpa is quite tiny
>
> are you running example on full example dataset or just the small one
> which is shipped within pymvpa?

I'm running with the full example dataset. The small one seems to
work. I will buy more RAM and try again with the larger set.

>> 2. If I ever manage to get the dataset imported, I'd like to run a PCA
>> analysis, then write out the component voxels again to a NiFTI-1
>> formatted file. I can't find any information on writing out data
>> again, except as a Python object. Is this even possible? If not, I'd
>> appreciate any nudges towards a solution.
>
> look at map2Nifti -- it can digest a numpy array of proper
> dimensionality (PCA components or projections, depending which way you
> do it) should have a matching dimension
>

Excellent thank you. I will try that.