Dear Amit,
> I was wondering if anybody has available, or could recommend, an analysis
> package (preferrably in Matlab) for fMRI image data that uses support-vector
> machine or similar classification methods for determining patterns of brain
> activity that maximally differentiate between conditions or between groups
> in an fMRI experiment...either using time course data or first level
> contrast images.
>
> There have been a few papers recently, including several in neuroimage, but
> it seems that the packages that have been cited have been developed for
> general use (ie not specifically with fMRI data and images in mind), and
> have thus been much adapted by the authors.
Our "Lyngby" Matlab toolbox contains an artificial neural network
classifier that has been applied on fMRI data. The toolbox is available
from:
http://hendrix.imm.dtu.dk/software/lyngby/
The algorithm used for one of the analyses in the following publication:
Nick Lange, Plurality and Resemblance in fMRI Data Analysis, NeuroImage,
10(3):282-303, 1999. http://dx.doi.org/10.1006/nimg.1999.0472
The approach is described in more detail in the manual and in my old
master thesis: http://hendrix.imm.dtu.dk/software/lyngby/manuals.html
Besides working as a classifier the method also returns a so-called
"saliency map" with indication of which voxels were important for the
classification.
best regards
Finn
____________________________________________________________________
Finn Aarup Nielsen, IMM DTU, Denmark
Lundbeck Foundation Center for Intergrated Molecular Brain Imaging
http://www.imm.dtu.dk/~fn/ http://nru.dk/staff/fnielsen/
____________________________________________________________________
|