[pymvpa] Searchlights and additional measures

Yaroslav Halchenko debian at onerussian.com
Sun Aug 23 18:31:28 UTC 2009

On Sun, 23 Aug 2009, John Clithero wrote:
> If I ran a searchlight analysis like the one I proposed, what would be
> the best way to save the peak performing (or other voxels of interest)
> voxel's coordinate/index and the index for the searchlight's
> neighboring voxels? I could then, I suppose, rerun the classifier on
> that searchlight ROI (without using the searchlight function) and
> harvest other properties (e.g., weights and confusion matrix) that I
> want.
> Again, many thanks for this, I am trying to learn quickly!
Sorry John, I do not quite understand what is the problem if any ;) you
just store your results in whatever structure you like (e.g. list or
ndarray), if you went through all indexes incrementally - you can simply
then use mapper.getInId to get coordinates... or just map all results
back into original 3D space and work there and get feature id for
interesting voxel with getOutId.

I would advise to store all performances through the whole brain/ROI
since it might give you better sense of 'by chance' distribution (all
those in the left tail of below-chance performances)
=------------------------------   /v\  ----------------------------=
Keep in touch                    // \\     (yoh@|www.)onerussian.com
Yaroslav Halchenko              /(   )\               ICQ#: 60653192
                   Linux User    ^^-^^    [175555]

More information about the Pkg-ExpPsy-PyMVPA mailing list