Dear Barbara and pickatlas users,
I have the same problem with pickatlas 3.0 and spm8 r4010.
I understand that the "apply masking" question of "results", is showing statitics done on ALL voxels of the brain but display only voxels inside the ROI (see the volume analysed at the bottom of your table)
If you use small volume, then the statistics are done only on the voxels inside the ROI, so the correction is less severe. It corresponds to what the "ROI analysis" done before...
That's right ?
Thank you
Isabelle
service de neurologie
INSERM U879
hopital Nord
42055 Saint-Etienne cedex 2
-------------------
Dear SPM users,
I assume there are no such things as stupid questions, so here we go :
After I encountered problems with the 2.4 pickatlas with the latest SPM8 update, I installed pickatlas 3.0 on a computer that runs a newer Matlab version. I computed ROIs and found two very nice clusters of activation.
Today, after reading Benjamin’s email about pickatlas 2.5 (Thank you!), I “upgraded” my pickatlas (running on Matlab 2006) and did the same ROI analysis again. Only this time, SPM came up with less and weaker activation than before.
I’m trying to understand how that is possible. Maybe I’m just overlooking something. I thought, in both cases (masking option of SPM that now includes the PA option, as well, and the ROI option in the 3.0 version) a mask was created under the same underlying assumptions, basically.
Does anyone know how this can happen? Benjamin, maybe you can shed light on this issue?
Thank you in advance,
Barbara
Dipl.-Psych. Barbara Rutter
Wissenschaftliche Mitarbeiterin
Abteilung für klinische Psychologie und Psychotherapie
Justus- Liebig-Universität Giessen
Tel.: +49 (0) 641/99-26085
Fax: +49 (0) 641/ 99-26099
Email: [log in to unmask]
-------------------------
Hello Barbara, Larissa, Desmond, and List.
Barbara has hit the nail on the head. With SPM8 Rev 4010 the SPM group has incorporated a method of ROI analysis into SPM. We are embracing this and have modified the version 2 branch of PickAtlas to work with this new method. We are releasing the WFU_PickAtlas 2.5 for SPM8 Rev 4010 or better. It may be downloaded from NITRC at http://www.nitrc.org/frs/?group_id=46 or http://fmri.wfubmc.edu/software/PickAtlas. When correctly installed, the "apply masking" question is now modified to "none","contrast","image", or "PickAtlas." (Actually: "none","con","img", or "PA" as screen real-estate is at a premium.) Selecting image will allow use of a preexisting image. Selecting PickAtlas will start the PickAtlas allowing on the fly ROI generation.
I am currently working on updates for PickAtlas 3.
Best,
Ben Wagner
ANSIR Lab
|