[pymvpa] digging for "the hidden treasure" - the PyMVPA atlas module

[Susanne Greiner]

Hi all,

Let's say you have done some MVPA and want to share your results with 
other researchers in terms of areas... why not give them a NIFTI? Maybe 
they do not use FSLview or they even co-registered to another common 
space, or simply do not like NIFTI ... not nice. So the only way is to 
refer to some atlas in order to make sure you are referring to the very 
same brain part.
The other way around, maybe they have found something interesting in a 
certain area with standard methods and also want a MVPA result, but do 
not give you ready to use masks in MNI-space, but just names of small 
brain areas. Again an atlas has to be consulted to launch your MVPA 
analysis straight away on that area of interest.
So one (not very comfortable way to do this) is to open FSL load the 
underlying atlas check the value of your area and have a look at the xml 
file to get the right name of that region. In the other case you start 
by the xml, open FSL, maybe create a mask which only contains the area 
with your value of interest and afterwards you even co-register it to 
your subject space...

The other day I came across the PyMVPA atlas module (which unfortunately 
is not yet documented). Looks like that could definitely simplify life 
in this kind of case. For example once you have loaded an atlas, there 
is the function label_voxel// which does all the looking up for you:

 >atlas = Atlas(filename="HarvardOxford-Cortical", resolution=2.0)
 >atlas.label_voxel([26,26,26],0)
 >{'labels': [[{'index': 39, 'label': 'Occipital Fusiform Gyrus', 
'prob': 19},
     {'index': 22, 'label': 'Lateral Occipital Cortex, inferior 
division', 'prob': 12}]],
     'voxel_queried': [26, 26, 26]}

'0.6.0~rc1' even seems to support using a standard FSL xml file for the 
atlas (filename) and a custom image file in your subject space (image_file).

So is anybody of you already using the atlas module a lot and can give 
me a hand with the following example:
- To label: searchlight results in subject space, lets call that dataset 
sl_map
- Preferably using Talairach (not my choice) from FSL
- the coordinate system of the atlas modul [26,26,26] would correspond 
to the sl_map.fa.voxel_indices!?
- what exactly does this levels option do?
- the other way round should be possible using the find option!? I tried 
searching for "Cortex", but as soon as there are multiple matches the 
option finishes in an error...

Any suggestions welcome! Thanks for all the work you did recently! Those 
multiple sa options are just great. Super comfortable for our kind of 
experiments!

LG

Susanne

-- 
Susanne Greiner, Dipl.-Phys. Univ.

ICT - Information and Communication Technolgy
International Doctoral School
NeuroInformatics Laboratory (NILab)
Via delle Regole 101, 38100 Mattarello (TN), ITALY

susanne.greiner at unitn.it | Phone +39 0361 882762 | Skype cattt84

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20110227/50c38f7b/attachment.htm>